Rochester Institute of Technology RIT Scholar Works

Theses Thesis/Dissertation Collections

1989 The introduction, demonstration and evaluation of a new typeface identification system J. Ames Parsons

Follow this and additional works at: http://scholarworks.rit.edu/theses

Recommended Citation Parsons, J. Ames, "The introduction, demonstration and evaluation of a new typeface identification system" (1989). Thesis. Rochester Institute of Technology. Accessed from

This Thesis is brought to you for free and open access by the Thesis/Dissertation Collections at RIT Scholar Works. It has been accepted for inclusion in Theses by an authorized administrator of RIT Scholar Works. For more information, please contact [email protected].

THE INTRODUCTION, DEMONSTRATION AND EVALUATION OF A NEW TYPEFACE IDENTIFICATION SYSTEM

by Ames Parsons

A thesis submitted in partial fulfillment of the

requirements for the degree of Master of Science in the School of Printing Management and Sciences in the College of Graphic Arts and Photography of the Rochester Institute of Technology

May 1989

Thesis Advisor: Professor Archibald D. Provan ACKNOWLEDGEMENTS

This thesis was a collaborative effort. Accordingly I wish to thank the following for their help:

Professor Archibald Provan for acting as my primary thesis advisor, for allowing me to virtually monopolize (for months) the Xerox Star system, for providing the fundamentals upon which this thesis is based, and for providing copious good cheer, patience, and support;

Professor Frank Cost for acting as my research advisor, and for designing and testing the TYPE-ID computer program;

Mr. Paul Swift, Director of the Graphic Arts Division of the RIT Research Corporation, for sponsoring this project, and for acting as my thesis advisor;

Mr. David Pankow, Curator of the Cary Library, for providing me with innumerable reference sources;

Professor Emery Schneider for allowing me and my helpers to use the typesetting equipment in the Phototypesetting Lab;

Professor Marie Freckleton for keeping the lab open;

Ms. Helga Birth, Librarian of the Technical and Education Center for the Graphic Arts, for her help and support;

Mr. John Winslow, President of Birdlow Associates, Inc. (Maplewood, NJ) and Jeff Wasilko for their typesetting help;

Tim & Maude Kent, Rob Kent, Mom and Dad, James Hamilton, Mary Lou Bulger, Sean McElveney, Jason Rosenberg, and various printing graduate students for their help in testing the various generations of the TYPE-ID system;

gang" Rick Schmidle and Mark DuPre and "the at the T&E Center for their unbridled humor and wit; and

My parents for their input, steadfast encouragement and support. TABLE OF CONTENTS

LIST OF TABLES vi

ABSTRACT 1

CHAPTERS

1. Introduction 1

Notes 3

2. Literature Search 4

Notes 8

3. Theoretical Basis of Study 9

The Panose Problem 9 The Pass Option 10 The Null Digit 11

4. Hypothesis 13

5. Definitions 14

6. Methodology 16

m 7. Results and Conclusions 18

Conclusion 1 18 Conclusion 2 19 Conclusion 3 20 Conclusion 4 21 Conclusion 5 24 Conclusion 6 25 Conclusion 7 27 Conclusion 8 28 Conclusion 9 28 Conclusion 10 29 Conclusion 11 29 Conclusion 12 29

8. Summary 31

9. Suggestions for Further Research 33

Suggestion 1 33 Suggestion 2 33 Suggestion 3 36 Suggestion 4 39 Notes 40

BIBLIOGRAPHY 41

APPENDICES

A. The Panose Number Generator 44

B. The Panose Classification Index 60

C. Samples of Panose Specimen Pages 64

D. TYPE-ID Profile Generator 71

E. Typeface Specimens 149

IV F. Data 201

G. Typeface Nomenclature 216

Notes 223

H. Typeface Classification Systems 224

Theodore Low De Vinne (1900) 225 F. Thibaudeau (1903) 225 Legros & Grant (1922) 226 Beatrice Warde (1935) 227 Johnson & Berry (1953) 227 Maximilien Vox (1954) 228

Deutsche Industrie Norman (DIN) System (1959) . . 228 ATypI (1961) 230 The British Standards System (1965) 231 Alexander Lawson (1971) 231 Marshall Lee (1979) 232 Rookledge & Perfect (1983) 232 Martin Solomon (1986) 234 Archibald Provan (1989) 234 Notes 240

I. Typeface Identification Systems 241

The gal Method 242 Rookledge's Earmark Tables 244 The 100-ems Method 246 The Panose System 248 Notes 249

J. TYPE-ID Designs 250 LIST OF TABLES

TABLE 1

Collected Data 202

TABLE 2

Average Search Yield and Probability of Success for Each Sequence 211

TABLE 3

Average Search Yield and Probability of Success for Any Search Based on One Sequence 212

TABLE 4

Probability of Success for Searches Based on One or More Sequences 213

TABLE 5

Average Search Yields for Searches Based on One or More Sequences 214

TABLE 6

Average Number of Sequences Needed to Narrow the List to One and the Probability of Success with TYPE-ID 215

VI ABSTRACT

Graphic communicators, including typographers, designers, and descriptive bibliographers, are often called upon to identify typefaces. Since there are thousands of typefaces available in the marketplace today, it is unlikely that any one designer, typographer or bibliographer can name them all at sight.

Accordingly, there is a need for a typeface identification system. The purpose of this paper is to introduce, demonstrate and evaluate such a system.

An identification system called TYPE-ID was developed and tested. Two new

option" digit" principles the "pass and the "null were introduced. A sequence of questions was written for a number of different characters. An

"database" index or of 50 typefaces was created. The sequences and database were incorporated into a computer program. Subjects tested the system 1000 times.

Based on the results, it appears that the system works. For searches of the database based on one sequence, the system proved to be 97.8% accurate. Chapter 1

INTRODUCTION

"identify" Graphic communicators are often called upon to (ie., determine the specific commercial name of) typefaces. To wit: (1) the typesetter who

"match" has been instructed by his client to a particular specimen of type must first determine the name of the typeface in which the specimen was set; (2) the designer who has happened upon a sample of a new or unfamiliar typeface (while perusing a periodical, perhaps) must first determine, before he can use the typeface in his next layout, the name of the typeface in which the sample was set; or (3) the descriptive bibliographer! who wishes to describe the type of a particular book, must, among other things, determine the name of the typeface in which the book was set.

There are thousands of typefaces available in the marketplace today. It is unlikely that any one designer, typographer or bibliographer can identify them all. Indeed, Geoffrey Dowding writes in The History of Printing Types:2

It is, of course, no part of the typographer's job to recognize and name at sight any specimen of typeface placed before him. Such is an impossible task and would only be demanded by one ignorant of the subject. 3

task" Because it is an "impossible to identify every specimen of type at sight there is a need for a typeface identification system. Such a system would enable its user, whether beginner or expert, to determine, accurately, the commercial name of the typeface in which any given sample of type is set.

Accordingly, the purpose of this paper is to introduce, demonstrate, and

evaluate the first step in the development of a new typeface identification

system. Chapter 1

NOTES

1. G. Thomas Tanselle suggests in his article entitled "The Identification of Description" Type Faces in Bibliographic (The Journal of Typographic Research, October 1967, p. 428) that a "precise description of the type used in a book is a proper part of the total bibliographic description of tnat book." Should this become the prevailing view, there would be a great demand for a typeface identification system. Particularly from those selected to identify the type in all the books in the Library of Congress.

2. Geoffrey Dowding, The History of Printing Types. (London: Wace & Co., 1961). xxiv.

3. Ibid. xxiv. Chapter 2

LITERATURE SEARCH!

The typeface identification system that has been developed in conjunction

System" with this thesis is based, in part, on the "Panose as presented by

Benjamin Bauermeister in The Panose System: A Manual of Comparative

Typography2 The specifics of The Panose System are discussed below.3

In order for the user of the Panose System to identify a typeface, he must

number." determine its "seven-digit classification This is accomplished by

generator." using the the supplied "number

The number generator is essentially a series of 7 questions. Each question

"generates" one digit.

The first digit of the seven-digit number the user must generate corresponds

to serif style. Accordingly, question 1 asks:

QUESTION #1:

What kind of serif structure does your typeface have?

1. cove serif 2. square serif 3. square cove serif 4. thin-line serif 5. exaggerated serif 6. triangle serif 7. square normal-end sans-serif 8. square perpendicular-end sans-serif 9. flared-end sans-serif 0. Rounded-end sans-serif If the user should determine, after studying the definitions of the above terms supplied in the introductory sections of the Panose system (see

serifs" Appendix A) that his unidentified typeface has "cove (answer 1

"1." above), then the first digit of his seven-digit number is If the user

serifs" should determine that his unidentified typeface has "square (answer

"2." 2) then the first digit of his seven-digit number is In short, the number

seven- corresponding to the answer selected becomes the first digit of the

digit number.

"proportion." The second digit the user must generate corresponds to

Accordingly, question 2 asks:

QUESTION #2:

What kind of proportion does your typeface have?

1. old style 2. modern 3. even width 4. expanded 5. condensed 6. monospaced

style" If the user should determine that his unidentified typeface has "old

proportion (answer 1) then the second digit of the seven-digit number is

"1." If the user should determine that his unidentified typeface has

"monospaced" proportion (answer 6) then the second digit of the seven- "6." digit number is In short, the number corresponding to the answer selected is the second digit of the seven-digit number.

This number-generating procedure is continued until all seven digits are generated.

index" Once the seven-digit number has been generated, the "classification

(see Appendix B) comes into play. If, for example the user has generated the number 1133122, he will find, in the numerically ordered classification index

"Baskerville." that 1133122 corresponds to Therefore, the user may conclude, if he has used the number generator correctly, that the name of

"Baskerville." his previously unidentified typeface is

To test the Panose system, two subjects were asked to try to identify 5 typefaces each. Both subjects concluded that the system is useless and gave

up in frustration.

The principal problem with the system is that the questions used to generate the seven-digit number are difficult to answer. Indeed, some questions are impossible4 to answer with certainty for some typefaces.

For example, the following problems were reported, among others, by test subjects attempting to answer the first question5: (1) the typeface

serifs" "Baskerville," which according to the classification index has "cove

"thin-line" "triangular" appeared to the users to have or serifs; (2) The Gothic," typeface "Serif which according to the classification index is a

sans-serif" "thin-line" "flared-end typeface, appeared to be a seriffed

Modern," typeface; (3) the typeface "Bernhard which according to the

"thin-line" classification index is a serif, appeared to have "triangular serifs";

Style" The typeface "Goudy Old has, according to the classification index,

serifs," "Schneidler," "cove but whose serifs resemble Goudy's, is said to

Gothic," have "exaggerated serifs"; and (5) the typeface "Letter which

"thin-line" appears to be a sans serif face, is classified, inexplicably, as a serif face.

Similar difficulties were encountered with each Panose question. None could be answered with certainty for every typeface. Chapter 2

NOTES

1. See Appendix I for discussion of other typeface identification systems. None of these systems is particularly relevent to this discussion, however, and none of them work.

2. Benjamin Bauermeister, A Manual of Comparative Typography: The Panose System (New York: Van Nostrand Reinhold Co., 1988).

3. See Appendices A, B and C for full description of the Panose System. Appendix A contains the Panose Number Generator, Appendix B contains the Panose Classification Index, and Appendix C contains sample Panose specimen pages.

questions" 4. This statement that "some of the Panose system are

certainty" "impossible to answer with is impossible to prove (and indeed, "impossible" any statement which says that something is is itself impossible to prove including this statement). Accordingly, the author is hesitant to support his troublesome statement with even a single example because any such example would not disprove the existence, or possible existence, of someone who can answer the Panose questions with certainty. Perhaps the

answer" phrase "the questions are impossible to should be replaced with the phrase "in our opinion, after exhaustive testing and research, the questions are extremely difficult to answer with certainty for even the most typographers." experienced However, in the interest of word economy, the "impossible" word will continue to be used indiscriminately, and perhaps a bit recklessly, throughout the text.

5. The specific problems that are mentioned below need not be understood substantively. The important thing is that there were problems. However, those wishing a more substantive understanding of tne Panose questions and answers may refer to Appendix A where a Full description of the Panose Number Generator is presented. Appendix B where the Panose Classification Index is presented, and Appendix C where some sample Panose typeface specimens pages are presented. Furthermore, a number of the basic typographical terms used throughout this paper are defined in Appendix G. Chapter 3

THEORETICAL BASIS OF STUDY

The new typeface identification system that has been developed in conjunction with this thesis is based on the Panose System (discussed in

Chapter 2) but has some important modifications. Below is a discussion of these modifications.

The Panose Problem

The problem with the Panose system as reported in Chapter 2 is that:

The questions and answers used to generate the seven-digit classification number are difficult to answer. Indeed, some questions are nearly impossible to answer with certainty for some typefaces.

At first glance, one might think that the solution to this problem is to change the questions so that they can be answered with certainty. After investi gation, however, it was concluded that this is prohibitively difficult to do.

Typographic details are so subtle and varied that it is difficult, if not impos sible, to design a set of questions, or even a single question, that is appli cable to, and answerable for, every existing typeface. Further, what may be a

"perfect" question today may not be a perfect question tomorrow since new designs are constantly being produced (many of which will undoubtedly be conscious departures from existing designs). 10

The solution to the Panose problem, therefore, is not necessarily to drop or alter imperfect questions (although each question should be made as perfect as possible) but rather to make provisions for imperfect questions.

The Pass Option

option" A system which incorporates the "pass is one which allows the user

"pass," to (ie., not answer) questions he cannot answer with certainty.

option" The Panose system does not have a "pass its user is required to

"incorrectly" answer every question. If he answers (ie., differently from the compiler of the classification index), his search of the classification index will end in failure.

For example, the user of Panose who has generated the number 2331221 for

"A" Typeface may have had difficulty in answering the first question.

Therefore, the first digit, a 2, may not have been generated with certainty. If

"3" the compiler of the index had previously generated a for the first digit,

Typeface A would be listed in the classification index as: 3331221. In this

classification index case, the user's search of the would yield either (1) no typefaces, or (2) a list of one or more typefaces which does not include the name of his unidentified typeface. 11

If, on the other hand, the user were allowed to pass on question 1, he would generate the following seven-digit number for Typeface A: P331221,

"?" where equals "pass".

Now, if the classification index is stored in a computer, the user can perform

"search" a of the index for the list of typefaces that have the same last six digits as his unidentified typeface. The first digit (where the user passed) would be ignored. Such a search would generate a list of typefaces that

does include the name of the user's unidentified typeface despite the fact that the user was unable to answer the first question.

The Null Digit

digit" A system which incorporates the "null is one which allows the

"pass" compiler of the index to on those questions he is unable to answer

with certainty.

If the compiler has generated the following number for Typeface "B":

3331221, but had difficulty in answering the second question, the second

determined with certainty. If user should digit, a 3, may not have been the

"2" assign a to the second digit, his search of the index would yield either

of typefaces which does not (1) no typefaces or (2) a list include his

unidentified typeface. 12

digit" If the compiler were instead to enter a "null as the second digit, the profile of Typeface B as entered in the computer database would be:

"#" digit." 3#31221, where equals "null The null digit is essentially an "all

above" of the response, ie., it matches any number input by the user for that digit. Therefore any of the following seven-digit numbers input by the user will be considered to match Typeface B:

3131221 3431221 3731221 3031221

3231221 3531221 3831221 3P31221

3331221 3631221 3931221 PPPPPPP

The location of these null digits can be determined by testing. For example, each question for every typeface to be entered into the index could be answered by a number of compilers. Where the compilers agree on an answer, the answer is entered into the index. Where they disagree, a null digit is entered. In this way, all the questions that are likely to "cause

trouble" can be found and provided for in advance. 13

Chapter 4

HYPOTHESIS

Statement of Hypothesis: The prototype of the typeface identification system that has been developed in conjunction with this thesis will enable its user to generate a list of the specific commercial names of one or more typefaces which includes the name of the user's unidentified typeface provided that the unidentified typeface is one of the 50 typefaces listed and displayed in

Appendix E of this paper. 14

Chapter 5

DEFINITIONS

TYPE-ID The name of the prototype of the new typeface identification system presented in this paper.

FULL PROFILE - The 76-digit description of a typeface which is determined

by answering all the questions of the 76-question profile generator.

PROFILE GENERATOR The 76 questions used to generate the profile of a

typeface. These are presented in printed form in Appendix A and are also

entered into the computer program. These questions are divided into

sequences.

"series" SEQUENCE - A related of questions. Sequences have been written for the following characters: caps A, C, G, J, L, M, P, Q, R, T and W;

lowercase a, e, f and g; figures 1, 2, 3 and 4; and the period punctuation

mark.

SEQUENCE PROFILE - The numerical description of a typefaces which is

questions for a specific sequence and which determined by answering the is

"n" digits in length (where n = the number of questions in the sequence).

SEARCH ARGUMENT -- The series of one or more sequence profiles upon which a given search is based. 15

SEARCH To compare the search argument of an unidentified typeface to the profiles of all the typefaces in the database.

DATABASE ~ The portion of the computer program where the full profiles of specific typefaces are stored.

LIST The list of typefaces found by a search which satisfies the search argument of the unidentified typeface.

SEARCH YIELD - The number of typefaces in the list.

MATCH - A match has been found when the name of the user's

unidentified typeface appears on the list.

"match." SUCCESS -- Synonym of

USER -- Person who attempts to identify a typeface using the TYPE-ID

system.

COMPILER(S) -- Person(s) who create the TYPE-ID database.

COMPUTER PROGRAM - The software in which the TYPE-ID system is

stored. The specifics of the system are not discussed in this paper. 16

Chapter 6

METHODOLOGY

In order to test the hypothesis presented in Chapter 4, the following steps

were taken:

(1) A sequence of questions was written for each of the following characters:

caps A, C, G, J, L, M, P, Q, R, T, W; lowercase a, e, f, g; figures 1, 2, 3, 4; and

the period punctuation mark. The questions can be found in Appendix A.

These questions serve the same function as the Panose number-generating

"profile" questions as each is designed to generate one digit of the of a

typeface;

"database" (2) An index or of 50 typefaces was created. This database serves the same function as the Panose classification index. Here, however, each digit for each typeface entered into the database was determined by three compilers. Where they agreed on a digit, the digit was entered into the database. Where they disagreed, a null digit was entered into the database;

(3) The sequences and the database were incorporated into a computer pro gram which allows the identification process to be completed with expedi ency. Specifically, the computer program allows the user to: (a) select a sequence (the user cannot access individual questions, he must access an

"n" entire sequence); (b) generate a sequence profile (which is digits in

"n" length where is the number of questions in the sequence); (c) translate 17

the sequence profile, including passes, into a "search argument"; (d) search

"list" the database for the of typefaces that satisfy the search argument; and

(e) repeat the process with another sequence (or sequences) and

immediately cross-reference between them;

(4) Subjects tested the system 1000 times. For each test, a subject selected

an unidentified typeface sample and then: (a) selected a sequence and

answered its questions; (b) entered the sequence profile into the computer

which translated the sequence profile into a search argument; (c) initiated a

search of the database; (d) counted the number of typefaces in the list

generated by the computer (thus determining the "search yield"); and (e)

answered the question "was the name of the unidentified typeface included

in the list?";

(5) The subjects recorded their results on data collection sheets similar those

presented in Table 1, Appendix F; and

(6) From the data collected, the system was evaluated both mathematically

and 9). and intuitively (See chapters 7, 8, 18

Chapter 7

RESULTS AND CONCLUSIONS

The results of the experimentation performed in conjunction with this thesis are presented in Tables 1 through 6 in Appendix F. The conclusions presented below are based on these results.

Conclusion 1

50 searches of the database were performed with each of the 20 sequences

of which TYPE-ID is currently composed (the raw data as collected by the

test subjects can be found in Table 1).

For each of the 20 sequences, a probability of success (Pseq) was determined

which is equal to:

Pseq = S/N x100

"S" recorded subjects for the specific where is the number of successes by

column and sequence in question (these values may be found in Table 2, 5);

"N" is the number of searches performed for that sequence (taken from

Table 2, column 2).

of success of a search based on the A For example, the probability cap

sequence is determined as follows: 19

Pseq = S/N X100

PcapA = 48/50x100

= PcapA .96x100

PcapA = 96.

"96," The result, indicates that, on average, there is a 96% chance that the user of the TYPE-ID system will find the name of his unidentified typeface in the list generated by the computer provided that the search is based on only the cap A sequence.

The "PSeq" values for the 20 sequences tested are presented in the far right column of Table 2.

Conclusion 2

Since 50 searches were performed with each of the 20 sequences, there were, in all, (50 x 20) 1000 searches performed. Of these, 978 ended in success. There is, therefore, on average, a 97.8% (978/1000 x 100) chance that the user of the TYPE-ID system will find the name of his unidentified typeface in the list generated by the computer provided that the search is

based on only, but any, one sequence. Or, more succinctly: For any search

based on one sequence, there is a 97.8% probability of success. This can be defined numerically as: 20

Pt = 97.8

"P-i" = where the probability of success with any one-sequence search.

This value of 97.8 is encouraging in that It suggests that the system works with near-perfect accuracy for searches based on one sequence. But anything less than 100% accuracy is unacceptable since any degree of error in a search based on one sequence is compounded by searches based on more than one sequence. To illustrate, if the probability of success of a search based on one sequence is 97.8 then the probability of success of a

search based on two is (.978 x .978) x 100 which is 95.6. With a search based on three sequences the probability of success decreases further. (This is discussed further in conclusion 4).

The only way to solve this problem is to ensure that the probability of success for any search based on one sequence is 100%.

Conclusion 3

While the computer is able to perform searches of the database based on the search arguments of two or more sequences, this feature was not tested.

However, the probabilities of success for searches based on more than one sequence can be determined mathematically as follows:

= (Pseq/100 X Pseq/I00 X . . . Pseqn/100) X 1 00 Pseqi, seq2, . . seqn 21

"n" where is the number of sequences in the search.

To illustrate, the probability of success of a search based on the cap A, cap

C, and figure 2 sequences can be determined as follows:

= PcapA, capC, fig2 (PCapA/100 X PCapC/100 X Pfig2/100 ) X 100

= PcapA, capC, fig2 (96/100 x 100/100 x 92/100 ) x 100

= PcapA, capC, fig2 (-96 x 1.00 x .92 ) x 100

= PcapA, capC, fig2 (.88)x100

= PcapA, capC, fig2 88

"88," The result, indicates that, on average, there is an 88% chance that the name of the user's unidentified typeface will appear in the list generated by the computer provided that the search is based on only the cap A, cap C and figure 1 sequences.

Conclusion 4

In the preceding example, the probability of success for a specific combination of three sequences was determined. Next, the average

sequences is probability of success for any search based on three (P3) determined. It is equal to either: 22

(1) the average of all the specific probabilities of success for each of the

([20 x 19 x 18] + [1x2x3] = 1140) 1140 combinations of three sequences, that is

P3 = (Pcombol + Pcombo2 + Pcombo3 + Pcombo1140) "^ 1140

which means "the average probability of success with any combination of 3

sequences is equal to the sum of the probabilities of all 1140 possible

combinations divided by 1140 (an equation which would be time-consuming to solve); or

(2) the following equation:

P3 = (P-|/100)3x 100

"P3" "3" Where is the average probability of success based on sequences;

"P-i" and is the average probability of success for one sequence.

uses 3 sequences on For example, any search of the database that will,

average, have the following probability of success:

P3 = (97.8/100)3 x 100

P3 = (.978)3 x 100

= P3 .93 x 100

P3 = 93% 23

This means that, on average, there is a 93% chance that the user of the

TYPE-ID system will find the name of his unidentified typeface in the list generated by the computer provided that the search is based on any, but

only, three sequences.

The basic form of the above equation can be used to calculate the average

probability of success for any search based on any number of sequences.

Accordingly, here is the basic form:

Pn = (P-|/100)n x 100

"Pn" "n" where is the average probability of success based on sequences;

"Pi" "n" is the number of sequences in the search; and is the probability of

success for one sequence.

search on 5 For example, the average probability of success for any based

sequences is equal to:

P5 = (97.8/100)5 x 100

P5 = (.978)5 x 100

= P5 .89 x 100

P5 = 89% 24

"Pn" The values for each sequence are presented in Table 4. Of particular importance is the fact that the P values decrease with each added sequence.

In other words, a search based on 3 sequences has a greater probability of success, and thus a greater P value (93.54), than a search based on 4 (91.45).

The reason for this can be explained as follows: Since the probability of

success for a search based on one sequence (Pi) is less than 100% (97.8)

and since the probability of success with n sequences is equal to Pi /100 to

the nth power, the value of Pn must decrease as the value of n increases. Pn

is, therefore, indirectly related to n. Because of this, the TYPE-ID system

becomes less accurate with each added sequence. It is therefore of

fundamental importance to improve TYPE-ID to the point where Pi = 100%.

Conclusion 5

The average search yield (Yseq) for each sequence is presented in Table 2.

Each Yseq value was determined by solving the following equation:

Yseq= TSeq/Nseq

"Tseq" "Yseq" where is the average search yield of a given sequence; is the

"Nseq" number of total yield of all trials of a given sequence; and is the total

searches made with the given sequence.

yield of the G sequence is equal to: For example, the average cap 25

YcapC Tcapc/NcapC

Ycapc = 977/50

YcapC = 19.5

This means that, on average, a search of the database based on only the cap

C sequence will yield a list of 19.5 typefaces.

Conclusion 6

The average yield for any search based on any one sequence can be

determined with the following equation:

Yi= T1/N1

"Y-|" where is the average search yield for any search based on one

"T-|" sequence, is the sum of the search yields for every trial made with one

"Ni" sequence; and is the total number of trials made with one sequence.

The equation may be solved as follows:

Yi= T1/N1

Y-|= 26396/1000

Yi= 26.4 26

For simplification, the average search yield for any search based on one sequence shall, hereafter, be defined as:

Yi = 26.

This means that, on average, the list of typefaces generated by the computer

after its search of the database contains the names of 26 typefaces. From the

"Yi" factor" value a "reduction may be determined as follows:

R = Y-i/B

"R" where is the reduction factor; "Yy is the "average yield for any search

sequence" "B" based on one and is the "total number of typefaces in the

database." The equation may be solved as follows:

R = 26/50

= R .52

This means that, on average, the list generated by the computer after a

search based on one sequence will contain 52% of the total number of

typefaces in the database. With this figure the yield for a search based on

sequences can be determined as follows: any number of

Yn = (R)n x B 27

"Yn" "n" "R" where is the yield for a search based on sequences.; is the

"w" reduction factor, is the total number of sequences used in the search,

"B" and is the total number of typefaces in the database. For example, the yield for a search based on 4 sequences can be determined as follows:

Y4 = (.52)4 x 50

Y4 =.073 x50

Y4 = 3.65

"3.65," The answer, means that, on average, a search of the database with 4 sequences will yield a list of 3.65 typefaces.

The calculated yields from Yi to Y20 are presented in Table 5.

Conclusion 7

The following is taken from Table 5:

Y6 = 1.

This means that, on average, the yield of a search performed with 6

order user of sequences is one typeface. Therefore, in for the the TYPE-ID

must on a system to narrow the list to one typeface he perform, average,

search of the database with 6 sequences. 28

Conclusion 8

If it is assumed that the user will continue to search the database until he has narrowed the list to one, and since the average number of searches needed to narrow the list to one is 6, the overall probability of success is equal to:

^type-id *

Where "P-^p^io- is the overall probability of success of TYPE-ID and P6 is the probability of success of a search based on 6 sequences. This equation may be solved as follows:

^type-id 6

= pr 7u/ -J54 TYPE-ID

This figure, 87.54, indicates that, on average, the TYPE-ID system, in its

search will be current form, is 87.54% accurate (assuming that the average based on 6 sequences).

Conclusion 9

= discussion are based Pi 97.8. All of the other P values calculated for this

system such that on P-|. It is therefore essential to improve the TYPE-ID Pi

will also 100. = 100. Once this is accomplished, all the other P values be 29

Conclusion 10

The imperfect probability of success rate may be attributable to the

digits" following: (1) the system may not contain sufficient "null which

means that the test subjects disagreed with the compilers of the database on

the answers to certain questions; (2) the test subjects made errors (such as

pushing the wrong buttons, answering the wrong questions) while inputting

their answers to their profile generating questions; or (3) the compiler made

errors while inputting the database.

Conclusion 11

Any sequence with a probability of success of less than 100 must be

investigated. These sequences can be identified from Table 2. Each question

"recompiled," for each of these sequences must be that is, each question

must be answered by a number of compilers to see where they agree and

disagree. Since the number of compilers used initially was 3, and since this

was apparently an insufficient number to determine the location of all the

null digits, the recompilation process, should, perhaps, be carried out with 5

or 10 compilers.

Conclusion 12

The recompilation process may prove inconclusive, that is, no, or few, new

null digits may be needed. If this is the case than at least part of the 30

experimental error may be attributable to test-subject error. This is not unlikely as only two test subjects generated all the data (each identified 500 typefaces!). Test subject fatigue may have been a factor. Upon retesting the system, a larger number of test subjects should be employed to answer a fewer number of questions each. The resulting data would be a more accurate reflection of the TYPE-ID system and much less a reflection of test subject endurance. The system may in fact be more accurate than the results indicate. 31

Chapter 8

SUMMARY

The first order of business of this chapter is to answer this question: "is the

false?" hypothesis presented in Chapter 4 true or

of."

The answer is "true . . . sort

To review, the hypothesis statement is:

The prototype of the typeface identification system that has been developed in conjunction with this thesis will enable its user to generate a list of the specific commercial names of one or more typefaces which includes the name of the user's unidentified typeface provided that the unidentified typeface is one of the 50 typefaces listed and displayed in Appendix B of this paper.

On average, 97.8% of the searches performed with one sequence yielded a list of typefaces which did include the name of the user's unidentified typeface (or, numerically, Pi = 97.8). The user who stops his search after one sequence can be reasonably certain that the name of his unidentified typeface is included on the list. Therefore, the hypothesis is basically true for searches based on one sequence. But, as shown in Table 4 and explained in

Chapter 6, the probability of success decreases when more sequences are added to the search. P20, for instance, is a miserable 64%. 32

The solution to this problem as discussed in chapter 7, may be two fold: (1)

more null digits must be incorporated into the system and (2) the procedure

for testing must be improved.

While the system is not perfect, in its current state, it is a considerable

improvement on the Panose system, and once the bugs are worked out, it

may very well work with 100% accuracy. 33

Chapter 9

SUGGESTIONS FOR FURTHER RESEARCH

The prototype of the typeface identification system developed in con junction with this thesis is not a comprehensive system. Its database is composed of only 50 typefaces and all of these are members of the same

"class" "seriffed" category or of type (Indeed, they may all be classified as typefaces). The purpose of this chapter is suggest some of the steps that must be taken to develop a comprehensive typeface identification system.

Suggestion 1

"bugs" Once the have been worked out of the TYPE-ID prototype, the database can be expanded to include all seriffed typefaces. If this proves successful, other profile generators and databases for other categories of

result: in finished be type may be developed. The TYPE-ID, its form, may

"branches" composed of several different each of which functions as a separate system and each of which is designed to identify the typefaces of a

"classification" specific category or of type.

Suggestion 2

If TYPE-ID is to be divided into classes, one question that must be answered

is this: What classification system should be adopted? 34

In Appendix H of this thesis are presented a number of typeface classification systems. From studying these systems the following conclusions were drawn (specific examples are taken from Appendix H or from the

sources cited in Appendix H):

(1) There is general disagreement as to what the major and minor categories of typefaces should be. For example: Some "Italic" treat typefaces as a separate category while most others Serifs" do not; One system does not include the subclass "Slab "Romans" under the broad category while most others do; One "Blackletters" includes in the miscellaneous grouping "Decorative" while many other have a separate category for Blackletters; Bold faces are treated separately by some and not by others; Display faces are divided into subclasses by some but not by others; and some have a separate category for "Private Types" Press while most others do not..

(2) There is general disagreement as to which typefaces belong "A" to which group; that is, a group in classification may have "B" the same label as a group in classification but their elements "Antique" may not be the same. For example: Thibaudeau's class "Antique" contains sans serif faces while the of DeVinne, Legros and Grant, and Martin Solomon contain only serif faces; "Gothic" DeVinne's class contains sans serif faces while "Gothic" Solomon's class contains blackletters.

(3) There is disagreement as to what the base of typographical classification should be. Some believe that a historical base is best; other believe a morphological base is best. But most "Roman" systems are a mix of the two. The ubiquitous category

serif," is historically based while such categories as "sans "slab "decorative" serif," and are morphologically based.

The point here is that care must be taken in the selection of a classification

system on which to base the TYPE-ID identification system. The system must

user of TYPE-ID can and be sufficiently clear that the quickly easily

-- determine to which class his unidentified typeface belongs a feat which

accomplish since typographical experts have been may be difficult to 35

debating the subject of typeface classification for decades (if not centuries)

and no system has yet gained universal support.

The beauty of the TYPE-ID identification system is its provision for

uncertainty. The user or compiler who is unable to answer one of profile

"pass." generating questions with certainty may Unfortunately, the user can

not pass on the classification question.

For example, if separate profile generators and databases are created for

"generic" each of the following classes of type

Blackletter

Uncial

Serifs

Sans Serif

Scripts

Decorative,

the first thing the user must do, before he can begin using any of the profile

which of the above classes his unidentified generators, is determine to

typeface belongs.

Roman Italic" difficulty. If his unidentified typeface is "Times he may have

"serif" Times Roman Italic has serifs and so might be classified as a face. But

and so might be classified as a script it also resembles cursive handwriting 36

face. Furthermore, depending on its use, it may also appear to be a

"decorative" face.

If this were a profile generating question, the user could pass. But here, he must answer to go any further. If he should answer incorrectly, he may spend time in the wrong branch of the system before he realizes his error.

This problem is not restricted to Times Roman Italic. There are probably hundreds, if not thousands, of typefaces that could be similarly misclassified

(especially by typographical novices).

The solution to this classification problem, then, is not to search for perfect class names, but to provide, as the profile generator does, for uncertainty.

This might be accomplished in one or both of the following ways: (1)

Typefaces which are likely to be misclassified may be entered into the databases of more than one class (for example, Times Roman Italic might be

"serif" "script" included in both the and databases) and (2) The classification question might instead be a series of questions which could have a profile generator and database of its own. The user would use the classification

profile generator to determine which identification profile generator to use.

Suggestion 3

The compliers of the various databases of which TYPE-ID will eventually be composed must make certain that they enter the correct names of their 37

typefaces. The most obvious mistake a compiler could make would be to

"Baskerville" "Bodoni." label profile A as when it is actually But, there are other, more subtle mistakes that could be made.

The commercial names of typefaces can be confusing and misleading.

Indeed problems have plagued the naming of individual typeface designs

since the incunabula days of printing. Geoffrey Dowding addresses this

problem in "Typefaces: A Plea for a Rational Terminology":

If one can criticize the founders for failing to agree among themselves on a less muddling system of terminology for their types, individual printers have likewise, without douot, further confused matters by the thoroughly arbitrary and reprehensible procedure of inventing names for the types in their cases. Were these invented names confined to the printer's own works all might be well, but as most printers issue type specimen sheets sooner or later these new names are broadcast and help to perplex many.1

example" Dowding gives what he calls a "flagrant of this name-changing

practice. A specimen book was published, sometime before World War II in

"Sun" Italic." which was listed a typeface called with its corresponding "Sun

It was later discovered that the specimens shown in the book were pirated

Roman" Italic." specimens of Monotype "Centaur and Monotype "Arrighi

change the but he also juxtaposed a roman Not only did the printer names,

together.2 and italic face that were not designed to be used

practice has been replaced a new menace. Today, this name-changing by

Alexander Lawson and Archibald Provan allude to this in their article entitled

"Commonly Used Typeface Names'^: 38

With the rapid expansion of photographic methods in typeset ting, there has also come an explosion of new typefaces readily produced by photographic means. A great many of these type faces, however, are simply adaptations of the most preferred basic types, altered to work with a specific equipment manu facturer's machinery and marketed under a new name. The resulting proliferation of typeface names has caused confusion among designers and editors.4

There are many examples of typefaces that have been slightly altered and

then sold under new names. Below are some examples:

"Baskerville" has been redesigned and sold under such names as "New

Baskerville," "Baskerline," "Beaumont." and

"Geneva," "Helios," "Helvetica" "Claro," "Corvus," has look-alikes named

"Megaron," "Newton," "Vega." and

"Elegante," "Palatino" "Andover," Hermann Zapf's has been marketed as

"Patina." "Melior" "Palladium," "Pontiac," and His has been marketed as

"Lyra," "Mallard," "Medallion," "Melier," "Uranus," "Ballardvale," "Hanover,"

"Optima" "Chelmsford," "Ventura." been marketed as and And, his has

"Orleans," "Ursa," "Zenith." "Musica," "Oracle," and

of other such examples. It should suffice to There are a veritable plethora

commercial names of a great typefaces are the say, however, that the many

confusion the typographic and that the sources of constant in industry,

should exercise caution. compilers of the TYPE-ID database 39

Suggestion 4

equipment" The discussion thus far has focused on the "standard of the

equipment" TYPE-ID system. There is, however, some "optional that might prove useful.

For example, after the user has tentatively determined the name of his unidentified typeface he might like some other information as a final check such as; (1) the name of the designer of the typeface, (2) the date the typeface was designed, (3) the country or countries in which the typeface is, or was, marketed, and (4) a list of typefaces which are similar in design or

"substitutable" for the unidentified typeface. Such information could easily be stored in a computer database and retrieved at the touch of a button. 40

Chapter 9

NOTES

Terminology," 1. Geoffrey Dowding, "Typefaces: A Plea for a Rational Typographica #4 (1926). 10.

2. Ibid. 11.

3. Alexander Lawson and Archibald Provan, "Commonly Used Typeface Names," appearing in Janet Field, Sr. Ed. Graphic Arts Manual (New York: Arno Press, 1980) 167-68.

4. Ibid. 167. 41

BIBLIOGRAPHY

Avis, F. C. Type Face Terminology. London: Glenview Press, 1965.

Basic Terminology for the Study of Letter Forms. Department of Printing Administration, Carnegie Institute of Technology, Pittsburgh (date unknown).

Bauermeister, Benjamin. A Manual of Comparative Typography: The Panose System. New York: Van Nostrand Reinhold Co., 1988.

Biegeliesen, J. I. The ABC of Lettering. New York: Harper & Row, 1971.

Centennial's Typeldentifier. Lancaster, PA: Centennial Graphics, Inc., 1986.

Chappell, Warren. The Anatomy of Lettering. New York: Loring & Massey, 1935.

Chappell, Warren. A Short History of the Printed Word. New York: Arno Press, 1980.

Cox, C. H., Ill, B. A. Blesser & M. Eden. "The Graphical Context of Printed Characters," Visible Language (Autumn, 1978). 430-447.

Dowding, Geoffrey. The History of Printing Types. London: Wace & Co., 1961.

Terminology," Dowding, Geoffrey. "Typefaces: A Plea for a Rational Typographica #4 (1926): 9-13.

Ettenberg, Eugene M. Type for Books and Advertising. New York: D. Van Nostrand Co., Inc., 1947.

Field, Janet, Sr. Ed. Graphic Arts Manual. New York: Arno Press, 1980.

Gaskell, Philip. "A Nomenclature for the Letterforms of Roman Type. "Visible Language (Winter, 1976). 41-61. Reprinted from The Library 5th Series (March, 1974). 42-51.

Gray, Bill. Tips on Type. New York: Van Nostrand Reinhold Co., 1983.

Hewitt, Graily. Lettering. New York: Pentalic Corp., 1976.

Johnson, A. F. Encyclopedia of Typefaces. London: Blandford Press, 1953. 42

Karch, R. Randolph. "Identifying Scripts." Ninth Graphic Arts Production Yearbook. New York: 1950. 597-599.

Kok, K. A. "gal" Print in Britain. 430-432.

Labuz, Ronald. Typography & Typesetting. New York: Van Nostrand Reinhold Co., 1988.

Lawrence, George H. An Introduction to Plant Taxonomy. New York: The Macmillan Co., 1955.

Lawson, Alexander. Printing Types: An Introduction. Boston: Beacon Press. 1971.

Lee, Marshall. Bookmaking, 2nd Ed. New York: R. R. Bowker Co., 1979.

Legros, L. A. and Grant, J. C. Typographical Printing Surfaces: the Tech nology and Mechanism of their Production. London: Longmans, Green & Co. 1916.

Standards." "Mare's Nest of Type Litho Review (April 25, 1984). 25.

Typefaces." Mosley, James. "New Approaches to the Classification of The British Printer (March, 1960). 90-96.

Typefaces." "Maximilien Vox Devises a New Classification of Printing World (July 30, 1954). 120-122.

Classification." "Maximilien Vox Discusses his New Type Printing World (August 8, 1956). 141; 156.

Technology," McCamy, C. S. "International Standardization in Graphic Taga Proceeding, 1977. Rochester: Technical Association of the Graphic Arts, 1977. 81-89.

Moxon, Joseph. Mechanick Exercises on the Whole Art of Printing, 2nd Ed. London: Oxford University Press, 1962.

Muenscher, W. C. Keys to Woody Plants, 6th ed. rev. Ithaca, New York: Cornell University Press, 1950.

Typefaces." Noordzij, Gerrit. "Broken Scripts and the Classification of Journal of Typographic Research (Summer, 1970). 213-241.

Perfect, Christopher & Gordon Rookledge. Rookledges International Typefinder. New York: Frederic C. Beil, 1983.

1946. Simon, Oliver. Introduction to Typography. London: Faber & Faber,

of New York: Simpson, George Gaylord. Principles Animal Taxonomy. Columbia University Press, 1961. 43

Solomon, Martin. The Art of Typography. New York: Watson-Guptill Publications, 1986.

Tanselle, G. Thomas. "The Identification of Type Faces in Bibliographical Description." The Journal of Typographic Research (October, 1967). 427-446.

Thompson, Tommy. How to Render Roman Letterforms. New York: American Studio Book, 1946.

Forms." Thorp, Joseph. "Towards a Nomenclature for Letter The Monotype Recorder (April/May 1931). 9-19.

Tracy, Walter. Letters of Credit, A View of Type Design. Boston: David R. Godine, 1986.

Tracy, Walter. "Type Design Classification." Visible Language (Winter, 1971). 59-66.

The Type Specimen Book. New York: Van Nostrand Reinhold Company, 1974.

Updike, Daniel Berkely. Printing Types: Their History, Forms and Use. New York: Dover Publications, Inc., 1980. Note: this publication is an unabridged replication of the second edition of the work as published by the Harvard University Press, Cambridge, Mass., in 1937.

White, Alex. How to Spec Type. New York: Watson-Guptill Publications, 1987. 44

Appendix A

THE PANOSE NUMBER GENERATOR 45

USING THE NUMBER GENERATOR

This book classifies a wide range of standard text choices are provided for how the serifs, or lack and display type styles. The following is an expla thereof, should appear. nation of the categories and features used to gen The selection of a serif style will greatly influence erate the seven-digit classification number by the appearance of the type style you choose. It is which the type styles in this book arc arranged. also the most difficult option set to work through. These options do not attempt to pinpoint the di The choices listed are very general; more minute mension, size, or weight of a particular typeface; variations arc usually available among the final al those variations arc shown on the display page or phabets found in the display section. It should also surrounding pages in the display section to which be noted that a serif style may change within a the classification number will guide you. single family as the character's weight increases. Many of the options provided in the different The PANOSE system uses the serif style of the me categories listed below arc applicable to only a few dium weight to classify the type. The ten choices type styles and arc rarely seen. These options arc are: followed by a bullet and should be avoided in the process of determining a classification number lor I. Cove serif the tout you desire, as you may not find an appro 2. Square serif priate type style with the number you specify, sim 1. Square cove serif ply because no such typeface exists. Usually these 4. Thin line serif options can be explored by using the cross-refer 5. Exaggerated seril ence chart once you have located a proper starting 6. Triangle serif page in th^j. display section. Remember, this book 7. Square normal end sans serif this 8. perpendicular end sans serif can quickly locate a typeface for you, but at Square point it cannot design one. Maybe next year. 9. Flared end sans serif 0. Rounded end sans serif

1. SERIF STYLE SERIF

A a A mk 1. Cove Serif

^^ ^^^ ^ LSent 2.SansSenl [ [

The first decision vouyou must make in selecting a type \ Cove Serif 1. Cove Sent

stvlc is whether or not the design you are looking serifs first option for a serif is the awe far for has feet, known as serifs, and what those The serif, by feet." Ten the most common of anv listed in this and it look like. S.ms serif means "without book, 46

is easy to distinguish. The cove serif is nothing 4. Thin Line Serif more than a line whose connection to the support ing leg has been softened by the addition of rounded coves to the inside corners. The cove serifs appear in a great variety of ways. If you are attempting to match a cove serif design, be certain that it fits these three criteria: it is signif icantly rounded at the inside corners otherwise it is a triangular serif; its tips come to a point, sharp \U54. Thin Line Serif 4. Thin Line Serif or dull, yet not flat; and the serif is a striking ele ment of the font's design. The cove serifs may be The thin line serif is the modern variation of the slightly cusped, dull, short, or long. If the serif is cove serif. The major difference between the two is very small, it might be classified as a flared sans that the thin line serif is not softened at the contact serif. point between the supporting arm and the serif. The thin line serif includes any serif style that is uncoved and light. In some instances the difference 2. Square Serif between a thin line serif and a square serif is very difficult to distinguish. In these cases the serif is classified according to its weight, the lighter being thin line and the heavier being square. Both of the serifs illustrated fall into the thin line

serif category. If the right-most example or the thin

line serif styles were any heavier or more flat-edged it would surely be a square serif. U-li?2. Square Serif 2. Square Serif

5. Exaggerated Serif * 3. Square Cove Serif

^^^^^^^^^_ L^^^k 5. Exaggerated Serif 5. Exaggerated Serif

3. Square Cove Serif 3. Square Cove Seril Because of the infinite variety of serif styles, the exaggerated is used to those The square serif and its companion, the square serif category classify styles known as styles that do not fall into the standard de cove, distinguish a group of type easily are the Egyptian family. The requirement for classifi scription categories. Included in this category serifs and the cation as a square serif is simple: a square serif the very heavy bone-shaped strongly serif. cusped designs. In addition to these there are must have at least two corners on each The types, softened also overemphasized and unclassifiable serif styles. square cove, as you might expect, is a these are that border on version of the strict square serif. Often fonts display fonts square classified as decorative. In comparison to other serif styles, the being identified. The serifs shown meet the requirements of serif (or slab serif) is the most easily may one of the other serif but because of their Identification of the square cove is moderately styles, of the con odd or exaggerated appearance, are classified more difficult because of the softening they a in this category. nection to the supporting leg. In classifying left and right square cove serif, be certain that the appearance; sides of the serif have a straight, flat similar to a otherwise the serif style might be more

exaggerated serif. cove serif or even to that of an 47

6. Triangle Serif style must either lie flat on the base line or meet it at 90 degrees. The square perpendicular leg ends, on the other hand, must always be at a 90-degree angle to their supporting arm. This is most obvious in diagonal letterforms, such as A. V, and some times M. The square perpendicular option is rarely used; thus it is followed by a bullet.

9. Flared End 6. Triangle Sent 6. Triangle Sent

Very few fonts incorporate the triangular serif. It is a sharp and awkward serif invented by the mind, not the eye. It is most easily seen in the typeface known as Maximal, a popular face, primarily by virtue of its triangular serifs. ll> Triangular serifs are very rare. Be certain, if you are matching an existing font, that the inside cor 9. Flared End 9 Flared End ners are absolutely flat; if they have any curve to The flared end style can be confused with a them, the font is likely to be classified as having easily cove serifs. small cove serif. Watch for two differences. First, most flared end style serifs start to widen much higher on the supporting leg than do small cove SANS SERIF serifs. Second, in a number of contemporary fonts, very minute serifs have been added to what tradi The remaining serif variations describe sans serif tionally were sans serif letterforms. In these cases, letter styles. The sans serif leg-end styles are much the serif is classified here as a flared end rather than easier to classify than the serif styles, mainly be a small cove because most of these "serited sans

serif" cause they have much less variation. faces, in keeping with other sans serif faces, have little or no contrast within the letterform. See 7. Square Normal End category 3 for a more in-depth description of con trast.

0. Rounded End

Ifilfi7 Square Normal End 7 Square Normal End

8. Square Perpendicular End

\!t0. Rounded End 0. Rounded End

The rounded end is very rare and should be used only to match an existing typeface. Category 0 is used when the leg end is round instead of flat. Moreover, additional fonts have been classified under this option if the overall or color of 8. Square 8. Square feeling the is rounded. In if a Perpendicular End Perpendicular End letters all, font has any rounded corners that are traditionally square in Sans serif letterforms have primarily two types of other fonts, it will be classified in this category. leg-end styles: the square normal and the square perpendicular. The leg ends of the square normal 48

2. PROPORTION consequently, the individual letter widths vary greatly. The geometric system by which the letters Proportion is not the shape of individual letters; of this style are constructed is based on the three rather, it is their relationship to each other. Pro primary letter shapes: the square, the circle, and portion is determined by such details as the geo the triangle. For example, the letter O is formed metric rules that are used to construct the letters, from a perfect circle, while the letter 5 is formed as well as overall specifications such as condensed from two circles placed one on top of another. This and expanded. results in the S being one-half the width of the O. The first three proportion categories are the The same situation occurs in the square shapes, as most difficult to identify because they are all sub demonstrated by the double-shaped letters F and sets of what we would normally call plain letters, . The letter E is formed by the stacking of two in that they are neither condensed or expanded. In squares, making it narrow in relation to the single- these options we see how the geometry of various shaped letters. Letters with triangular shapes have letterforms works to create a well-balanced alpha the same characteristic: the construction of the bet. double-triangle X is half the width of the single- triangle V. Type designs from the modem period lose their 1. Old Style strict geometry and become less varied in width. 7 \ The circular O becomes an oval, and the arms of the once narrow letter E are elongated to make it \ more compatible with the single-shape letterforms. Finally, there is the even width category, typified by more contemporary type styles, where letters have been stretched and compressed until they arc lESON ot all of nearly equal width. You can see the return

1 . Old Style Proportion the strict full-size geometric shapes in letters such as O and V, while the two-part letters continue to 2. Modern expand, as seen in the letters E and 5. Also in cluded in this category are the fonts that, while neither modern nor condensed, have full-form let ters that are slightly taller than they are wide, such as Eurostile and Univers.

4. Expanded

2. Modern Proportion 7 \

3. Even Width \ 7 \

4. Expanded Proportion

ESON Those letters with standard widths greater, than tin height of the uppercase letters are classified as c.\

idei' 3. Even Width Proportion painted. Expanded lettertorms cm be c.isiK tified by looking at the capital O. If it is a wuii characters haw followed an oval, the overall dimensions of the Letter designs throughout history have a style of great stahilin progression with re been expanded, creating amazingly straightforward width. changes. The type and gard to their proportional early style. The pro styles are generally known as old geometrical; portions of these letters are strictly 49

5. Condensed

5. Condensed Proportion

Condensed letters are the opposites of expanded letters. The condensed option is a little more diffi cult to judge because no single character can be used to identify it easily, as the O can for expanded fonts. Most often the font name itself will be iden

tified as condensed. In those instances where this is not the case, the uppercase O is again used for reference, although it is difficult to distinguish be tween a modern and a condensed O, both being ovals. The more exaggerated the oval, the more likelv it is that the font is condensed.

6. Monospaced (Typewriter) ?????D lESO NM 6. Monospaced Proportion

Unlike even width which has the appearance of even spacing between the letterforms, though the rules arc broken for narrow letters like / and wide

letters like M each letter in a monospaced font

has the same character width and spacing. Mono spaced fonts are also typically known as typewriter fonts because they mimic the letter widths and spacing created by an old-fashioned typewriter. With the advent of modern typesetting, many of these styles became available in various media from hot to digital type, and several were devel oped into fonts in their own right. Some of these

so-called typewriter fonts have dropped their true monospacing in favor of greater legibility; none theless, they are classified as monospace because of their typewritten appearance. 50

3. CONTRAST 3. Medium

Medium contrast is the most common option for Most letters stems of have and arms varying thick serifed fonts. Included in this selection are the myr the nesses; usually horizontal members are thinner iad of fonts that have good typographical contrast than the vertical. The Contrast category details the yet are not designed to draw attentton to their relationship between a letter's members, describing structural variety. the structural contrast between the heavy and the 4. Medium High light elements of the individual letterforms. Con trast does not have anything to do with the sur Most of the modern fonts have medium high con rounding whiteness of the page or the contrasts trast, with a visible contrast between the stems and among different letters; this category deals only arms of the letters. This contrast is accentuated by with contrasts in elements of a character. Be care the use of equally thin serifs. ful not to confuse Contrast with weight when ex 5. High amining options here. Once again, the uppercase O acts as a good reference character to determine Those fonts that use extreme contrast between the contrast of a given font. stems and arms to create a dramatic effect are the The choices you make in this Contrast category only styles classified as high contrast. will determine which set of options you should 6. None look at in the next category. Stroke Variation (4A) or Arm Type (4B). If you choose options 1 through If a type style has no variation in width between 5 in this category, go to 4A, Stroke Variation. If the stems and arms of its characters, it is said to be you choose option 6. go to 4B, Arm Type. a single-stroke letterform and is classified here as no contrast. Most of the fonts so classified 1. Low having will be sans serif styles, although many square serif Fonts that have low contrast arc generally those styles are also single stroke. Often when the weight senfed styles with only a very slight variation be of a font becomes heavy, contrast is increased to tween the weights of the stems and the arms. provide adequate space for the horizontal arms;

this increase in contrast also occurs in the rounded 2. Medium Low letterforms when they have been altered to increase Medium low styles show subtle variations in con legibility. Fonts arc classified as having no contrasr trast and maintain a very soft appearance. Many if the medium weight of the font is single stroke in of the Old Style fonts will be found in this option, form. then as will the optically refined sans serif fonts, such as If you have chosen option 6, no contrast, Optima and Souvenir Gothic. you will use category 4B to select your fourth digit. |0|0|0 Medium Contrast 1 . Low Constrast 2. Medium Low Contrast 3. |OIO|0

4. Medium High Contrast 5. High Contrast 51

4A. STROKE VARIATION I. Gradual with diagonal stress 2. Gradual with mixed stress

J. with vertical stress rden-~Gradual This category, used for double-stroke letters, 4. Gradual with horizontal stress tifies the speed at which a letter's stroke changes 5. Abrupt with vertical stress width and the direction of stress created by-that 6. Abrupt with horizontal stress change. The options for speed are gradual, abrupt, 7. Instant vertical stress and instant. They are coupled with the options for stress, diagonal, vertical, horizontal, and mixed. These two sets of options are combined and listed Option 2 provides a choice that is fairly rare. Dur below in their historical order, starting with typog ing what is known as the transitional period of raphy's calligraphic beginnings, demonstrated by type design, the stress of the letters changed from the diagonal stress and the gradual shifting of letter diagonal to vertical. Some alphabets designed dur weight, continuing through the transitional period ing this period changed the stress in only the up of mixed stresses, then to the modern period with percase; the lowercase retained its diagonal stress. its abrupt and strictly vertical stress, and finally the Thus, these faces have mixed stress. This option is more decorative stroke variations of horizontal not onlv seldom used, it is also often overlooked. and instant vertical.

With 1 . Gradual With 2. Gradual With 3. Gradual Diagonal Stress Mixed Stress Vertical Stress

4. Gradual With 5. Abrupt With 6. Abrupt With Horizontal Stress Vertical Stress Horizontal Stress

7. Instant Vertical Stress 52

4B. ARM TYPE concerns stem and arm style. The definitions are quite simple. Straight arms applies to those fonts To create some distinction among the various with arms (especially those of the uppercase A and single-stroke font styles, this category details two W) that remain straight and parallel. Some of the different aspects of letter design. Although most o bolder display fonts, which rely on the ability to these options arc directed toward sans serif letters bend their normally straight arms to compensate the few single-stroke serif designs that exist will be tor their unwieldy weight, fall into the hotved arms accommodated as well. The two variations covered option. The final option is for those fonts with here are Arm/Stem style and Round Letter Open concave stems and arms. These fonts arc easiest to ing style. Each of the two topics has three varia distinguish by evaluating the uppercase /, the sides tions; they are grouped into nine different options; of which will curve inward, rather than being per fectly parallel. 1. Straight arms, horizontal opening The second set of options defines the angle at 2. Bowed arms, horizontal opening which the opening of the uppercase C (and conse 3. Concave arms, horizontal opening quently the G and S) is set. See the illustration for 4. Straight arms, wedge opening an exaggerated display of horizontal, wedge, and 5. Bowed arms, wedge opening vertical openings. In the horizontal option, the 6. Concave arms, wedge opening opening is parallel. If either of the opening angles 7. Straight arms, vertical opening is not horizontal or vertical, then the font has a 8. Bowed arms, vertical opening wedge opening. Sometimes a wedge opening will 9. Concave arms, vertical opening actually be an inverse wedge. Most of the serifed fonts classifed as single stroke have a vertical open The first choice you must make in this category ing. |A|A|A

3. 1 . Straight Arms 2. Bowed Arms Concave Arms |C|C [C 3. Vertical 1 . Horizontal Opening 2. Wedge Opening Opening 53

5. LETTERFORM boxed option, in winch there is still only one con tact point per side, but the overall form of the

rounded letters is square. In the or The form of a letter is very important for establish flattened, the letter will contact the ing the mood and appearance of any style. Classi straight-sided, option, edge of its external parameters tor an extended fication is determined by evaluating the rounded letterforms, such as O and C, for six variations: length, thus flattening the sides ol the rounded shapes. The top and bottom of the letterform, I. Contact however, will still contact the dimensional limits at 2. Boxed one point. The rounded option is provided for 3. Flattened those forms that are primarily squares with dull 4. Rounded corners, contacting the parameters for extended 5. Off center lengths on all four sides. Similarly, the square 6. Square option is available for completely sharp, square-cornered letterforms. Off center forms are The first option, contact, is by far the most com predominantly asymmetrical forms. Note that a mon: the exterior outline of the form comes in contact style need not be a true circle: it can be an

contact with the visual parameters of the letter's oval without flat sides. dimensions at only four points. Similar is the

121 Contact IS2. Boxed II3. Flattened

ISIOI14. Rounded 5. Off center 6. Square 54

6. MIDLINE midline of the A considerably lower. This style is classified as high standard. The last two options, The Midline category encompasses two aspects of both rather decorative, pertain to those fonts with type design. The first option determines the posi out a variation in the midline of the A. These two tion of the midline throughout the alphabet. This options apply to fonts that have their midlines en option offers four choices. These four choices are tirely set in the middle, constant midline, and those combined with the second option, which defines that are entirely set low, constant low midline. whether or not the apexes of pointed letterforms, There are also a very few entirely high-set midline such as uppercase A, W, and M, are trimmed flat positions, which have been classified under the or remain sharp. high standard option. A trimmed apex is, it would seem, a very simple 1. Standard midline with trimmed apex concept to visualize: if the top of an uppercase A 2. Standard midline with pointed apex is flat, it is considered trimmed; if it is not flat, it is 3. High standard midline with trimmed apex considered pointed. Yet two difficulties arise. First, 4. High standard midline with pointed apex many sans serif fonts have only a very small flat 5. Constant midline with trimmed apex area, one that is hardly noticeable. Second, some 6. Constant midline with pointed apex serifed fonts with sharp apexes have right arms 7. Constant low midline with trimmed apex that overlap the left arm beneath and to the left of 8. Constant low midline with pointed apex the apex. (For example, see the serifed A in the serif style illustration on page ix.) Hence two Type designers have handled the midline posi more rules are added to the determination of the

tions in various ways. Standard midline position apex trim: to be considered trimmed, the flat area generally is centered slightly high throughout most at the top of the letter must be greater than half of the alphabet, with the exception of the upper the width of the wider arm to be considered case A, which is usually set equally low. A common trimmed, and to be considered pointed, the point variation on this standard is to place the midline at the top of the uppercase A must be clean and undisturbed to be considered pointed. noticeably higher than average and to drop the (A|A

1 . Trimmed Apex 2. Pointed Apex IA-E-H IA/E-rt

1 . Standard Midline 2. Standard High Midline IA--E-H- IA-E-14-

4, Constant Low Midline 3. Constant Midline 55

7. x HEIGHT less than SO percent of the uppercase height are considered small. Lowercase letters taller than 70

The x height of a typeface is the height beyond percent of the uppercase height are classified as which the bodies of lowercase letters do not ex targe. In many display faces, no lowercase is avail tend, generally measured as the height of the low able, and these are the alt-cap fonts of option 4. ercase X. The size relationship between uppercase This concludes the number-generating process. and lowercase letters varies considerably and is mi Make sure you have seven digits in your number, important factor in a style's readability as well as and then locate the seven-digit number in the Clas the overall color of a page of typed copy. The four sification Index. This index will provide a corre options here are based upon the relationship of the sponding display page number. If no index number lowercase letters to the uppercase. is similar to the classification you have chosen, re- check your number, making sure the digits are in 1. Small proper order. If they are correct and you still find 2. Standard no similar listing in the Classification Index. 3. Large choose the closest number available in the index 4. All caps and turn to the corresponding display page. It is often possible to use the cross-reference chart on Lowercase letters slightly larger than half the up the display page to find alternate styles of the given percase size are standard. Those lowercase letters type that better fit your needs. lx|(xl(x

1 . Small x Height 2. Standard x Height 3. Large x Height IX

4. All Caps 56

EXAMPLES

As mentioned earlier, PANOSE can be used in a cause many of the options in the Letterform variety of ways. This section gives two examples of category are uncommon, as indicated by the bul how to apply the PANOSE system. The first dem lets, you select option I, contact, to make certain onstrates a standard search for an appropriate type that you will have a variety ot styles from which to face, and the second shows how to use PANOSE choose. Now that you are starting to get a feel for to match an existing typeface. the overall appearance of the font you are select ing, choosing the last two digits will be much eas SEARCHING FOR A TYPEFACE ier. In the sixth category, Midline, you decide to follow the angular mood of the font and choose 2, Suppose that you know nothing about type styles standard midline with pointed apex. Finally, you and have been given the task of selecting a tont for select a size for the lowercase. Since you want to a flyer or advertisement; or better yet, you have have good legibility, you opt for a large x height been given the task of helping a client who knows and choose 3. nothing about type select a font. Your seven-digit classification number is there You feel that the font needs to be particularly fore 8 164123. When you turn to the Classification bold to catch the reader's eye, so you decide to Index, you find that the closest font to your num

look at some sans serif styles. As you ponder the ber is Syntax, classified as 81641 12 and displaved

arc- four sans serif options in the first category, you on page 213. When you flip to that page you decide to be just a little different, so you choose delighted to find that the midline position you had option 8, perpendicular sans serif, as your first originally specified is actually the position of the digit. The second digit is not as easy to select; you boldface font. In addition, you notice that this font know that you do not want a condensed or ex is available from a variety of suppliers, so you panded font, but beyond that you are not too cer should have no problem ordering it. tain which proportion scheme will work best. After a moment of hesitation, you decide on option 1, old style, as your second digit since it seems to MATCHING AN EXISTING have the most personality. Since most sans serif TYPEFACE font styles are single stroke, you will have a greater the no- In this second example we will deal with a variety of fonts to choose from by selecting usually select the problem. You have just been handed an contrast (single-stroke) option when you frustrating digit from the Contrast category. Therefore your advertisement that your predecessor created just was promoted and transferred to an third digit is a 6. Because you selected a 6 from the before she and de office across the country. You have been instructed Contrast category, you skip category 4A the ad so your bate which pair of options from 4B will work best. to match the type style from boss indi can use it to print some napkins a A review of the digits you have chosen so far matching for big

standard font. open house next month. The last time you did this cates that you are selecting a rather straight-armed it took six and the end of you You therefore decide to stay with a chore, hours, by it, to match had blisters on both thumbs and all of the fonts letter, and you choose a wedge opening cate the same and dif the perpendicular sans serif you selected in looked both exactly completely simple. Be at the same time. gory 1. The next choice is relatively ferent, 57

But now you have the PANOSE system, in standard in appearance, you choose a 1. The x which all of the similar-looking fonts are grouped height is easy to gauge, as the lowercase is almost together, with pointers showing how they differ. exactly 60 percent as tall as the uppercase. You jot With a sample of the font in front of you, it is easy down the last 2, and your digits arc complete. Now to pick the digits that correspond to the font you you must check if the faces match.

are to match. a trying The first digit is obviously 1 , In the Classification Index, you find no exact the cove serif option, which is unfortunately the match for your number 1323212. The closest most commonly used sent style available. Un number to yours is 1321 1 12, ITC Novarese. You daunted, you proceed onward. The next digit is a turn to that page and realize immediately that little tougher. Because your sample docs not have something is not quite right. The font on your sam enough letters on it, you cannot tell whether the ple is somehow more dynamic or diverse. You proportion is old style or even width. Taking a stab wonder if any font is just like the one displayed, in the dark you choose 3 for your second digit, only more classical. You check the cross-reference committing yourself to the even width option. The chart and see that if the second digit, for propor arbitrary nature of the Contrast category slows tion, were changed from 3, even width, to 1, old you down a bit; you decide to pencil in a 2 for style, another face would be a perfect match. You medium low contrast. You have much less doubt turn to the page listed in the chart and notice the in your mind about selecting a 3 for category 4A. new font, CG Bern, is too classical and too antique. Your sample definitely has a gradual shift in weight You realize that the problem arose when you ne as it moves from thick to thin around the perimeter glected to match your fourth digit for arm style, of the uppercase O, and the thinnest points of the gradual and vertical. Checking the cross-reference O are directly above one another, giving it a feeling chart again, you locate another similar font, and of vertical stress. The Letterform category is more this one, although it is not a perfect match, has a 3 challenging; it is unclear whether the rounded for the fourth digit. When you turn to that page, shapes are all perfect circles or slightly boxed, so you notice that there is a red line under the classi you ]ot down the number 2. For the sixth digit you fication number, meaning that there are several are again a bit uncertain because you do not have fonts with the same number. After comparing the

a sample of the letter A to judge the midline posi specimens with your sample, you find that Edel tion. Considering that the rest of the font has been weiss matches it exactly. 58

NUMBER GENERATOR

Typefaces in this book are cataloged by a number If you select a typeface with no contrast (6), select that is obtained by selecting options from the fol from category 4B. If you choose any other contrast lowing seven categories. Explanations of these op level (1-5), use category 4A. tions are on pages ix-xix; a review of these explanations may prove useful to anyone using the 4A. Stroke Variation (double-stroke letters) number-generating process. Keep track of your se 1 = Gradual with diagonal stress lected digits on a separate piece of paper. Digits are 2 Gradual with mixed stress

generated left to right. A classification number fol 3 = Gradual with vertical stress lowed by a red bullet indicates a rare classification 4 = Gradual with horizontal stress option and should be avoided, unless you are 5 = Abrupt with vertical stress

= trying to match an existing type. 6 Abrupt with horizontal stress 7 = Instant vertical stress

4B. Arm Type (single-stroke letters) 1. Serif Style 1 = Straight arms, horizontal opening 1 = Cove serif 2 = Bowed arms, horizontal opening 2 = Square serif 3 = Concave arms, horizontal opening 3 = Square cove serif 4 = Straight arms, wedge opening

= 4 = Thin line serif 5 Bowed arms, wedge opening

= 5 = Exaggerated serif 6 Concave arms, wedge opening

= 6 = Triangle serif 7 Straight arms, vertical opening

= 7 = Square normal end sans serif 8 Bowed arms, vertical opening

= 8 = Square perpendicular end sans serif 9 Concave arms, vertical opening 9 = Flared end sans serif 0 = Rounded end sans serif 5. Letterform 1 = Contact 2. Proportion = Boxed

= 1 = Old style Flattened

= - 2 = Modern Rounded

= 3 = Even width Off center

= 4 = Expanded Square 5 = Condensed 6 = Monospaced Midline 1 = Standard midline with trimmed apex 3. Contrast 2 = Standard midline with pointed apex = standard midline with trimmed apex 1 = Low 3 High = standard midline with pointed apex 2 = Medium low 4 High = midline with trimmed apex 3 = Medium 5 Constant = midline with pointed apex 4 = Medium high 6 Constant = Constant low midline with trimmed apex 5 = High 7 = Constant low midline with pointed apex 6 = None 8 59

x Height 1 = Small 2 = Standard 3 = Large 4 = All caps

M

S = Serif Style P = Proportion C = Contrast A = Stroke Variation/Arm Type and Opening F = Letterform

M = Midline Height and Apex X = x Height

Proceed to the Classification Index to find the cor responding display page number for your new clas sification number. 60

Appendix B

THE PANOSE CLASSIFICATION INDEX 61

CLASSIFICATION INDEX

1111-162 Brewer Text 1 1233-113 Romana 43 1113-113 ITC Weidemann 2 1233-122 Gazette 44 1113-133 ITC Usherwood 3 1235-112 Imprint 45 1121-111 Simoncini Garamond 4 1235-122 ITC Century 46 1121-112 CGBem 5 1235-123 ITC Isbell 47 1121-112 Garth Graphic 6 1245-112 Madison 48 1121-112 Sabon 7 1245-112 ITC Tiffany 49 1121-113 Administer 8 1245-113 ITC Modern No. 216 50 1121-121 Artcraft 9 1245-122 Bulmer 51 1121-121 Berling 10 1311-121 Cloister Old Style 52 1121-121 Eldorado 11 1321-112 ITC Novarese 53 1121-121 Stempel Garamond 12 1331-113 Adroit 54 1121-122 Goudy Old Style 13 1332-532 Windsor 55 1121-122 Simoncini Life 14 1333-212 ITC Zapf International 56 1121-132 Expert 15 1343-121 Lucian 57 1121-133 ITC Benguiat 16 1343-163 Independence 58 1121-142 Kennerly 17 1345-122 Devinne 59 1121-143 Musketeer 18 1521-133 ITC Benguiat Condensed 60 1121-152 Griffo 19 1531-123 ITC Garamond Condensed 61 1121-212 Poppl-Pontifex 20 1535-122 ITC Century Condensed 62 1121-221 Aldus 21 1545-122 Century Nova 63 1122-131 Collage 22 1555-122 Trooper 64 1123-122 Edelweiss 23 2113-112 Joanna 65 1123-122 Granjon 24 2121-112 Breughel 66 1123-122 Times Roman 25 2121-252 Cartier 67 1123-131 Cochin 26 2123-123 Candida 68 1123-142 Brighton 27 2167-112 Stymie 69 1123-162 Perpetua 28 2167-113 ITC Lubalin Graph 70 1131-112 ITC Galliard 29 2221-113 Vladimir 71 1131-123 ITC Garamond 30 2223-122 Fairmont 72 1133-122 Baskerville 31 2223-212 Stemple Schadow 73 1133-122 ITC New Baskerville 32 2233-112 Auriga 74 1133-122 New Caledonia 33 2235-123 Else 75 1133-122 Janson 34 2261-613 Geometric 76 1143-112 Caslon No. 540 35 2267-113 Glypha 77 1155-112 ITC Caslon No. 224 36 2311-113 Eternal Oldstyle 78 1164-242 Camelot 37 2323-133 ITC Leawood 79 1221-133 Caxton 38 2325-223 Antikva Margaret 80 1223-122 Primer 54 39 2361-112 A&S Gallatin 81 1231-112 Leamington 40 2364-112 Rockwell 82 1233-112 ITC Esprit 41 2367-112 Memphis 83 1233-112 New Aster 42 2367-112 Monty 84

247 62

2536-313 Playbill 85 4255-122 Bodoni 141 2536-313 Branding Iron 85 4321-111 Bernhard Modern 142 2561-423 City 86 4333-212 ITC Zapf Book 143 2567-112 Stymie Condensed 87 4357-113 Normande 144 2567-312 Memphis Condensed 88 4661-113 Letter Gothic 145 3113-132 Bramley 89 4667-112 Courier 146 3121-112 Palatino 90 4667-113 Typewriter Large Elite 147 3121-133 Belwe 91 4667-113 Typewriter 147 3121-212 Raleigh 92 4667-123 Prestige 148 3123-111 Cheltenham Old Style 93 5111-162 Schneidler 149 3123-113 Cardinal 94 5221-112 Worchester Round 150 3133-122 Caledonia 95 5231-113 Packer Text 151 3167-112 Beton 96 5311-132 Bryn Mawr 152 3213-112 ITC Italia 97 5323-123 Scenario 153 3213-212 Congress 98 5323-132 ITC Souvenir 154 3223-112 Century Old Style 99 5323-212 Hadriano 155 3223-112 Egyptienne Frutiger 100 5331-113 Cooper 156 3223-112 Textype 101 5523-133 Seagull 157 3223-113 ITC Cheltenham 102 5567-123 ITC American Typewriter 3223-122 Concert 103 Condensed 158 3223-123 Dominante 104 5667-123 ITC American Typewriter 159 3231-113 Plantin 105 6121-112 Trump Medieval 160 3233-112 Accolade 106 6133-122 ITC Veljovic 161 3233-112 Astro 107 6223-113 Maximal 162 3233-113 Criterion 108 6223-122 Versailles 163 3233-122 Century Schoolbook 109 6233-122 ITC Gamma 164 3235-122 Century Expanded 110 6313-113 Barcelona 165 3235-122 New Century Schoolbook 111 6323-112 ITC LSC Book 166 3235-123 Corona 112 6333-123 Meridien 167 3313-113 Quint 113 7123-122 Peignot 168 3313-132 ITCKorinna 114 7161-113 ITC Avant Garde Gothic 169 3313-213 Egyptian No. 505 115 7164-153 Martin Gothic 170 3323-112 Clarendon 116 7167-112 Frutiger 171 3323-112 Pasquale 117 7167-112 Gill Sans 172 3323-113 Bookman 118 7167-141 Spartan 173 3323-113 Excelsior 119 7167-142 Futura 174 3323-113 Impressum 120 7167-142 ITC Ronda 175 3323-113 Ionic 5 121 7167-253 ITC Eras 176 3323-123 Craw Clarendon 122 7213-213 Imago 177 3325-252 Melior 123 7245-113 Brittanic 178 3343-123 Claire News 124 7261-113 Uncia 179 3367-113 Helserif 125 7261-113 Video 180 3523-112 ITCCushing 126 7264-112 Monotype Grotesque 181 3533-113 ITC Cheltenham Condensed 127 7264-112 News Gothic 182 3545-212 Renault 128 7264-112 Trade Gothic 183 4121-111 Trajanus 129 7267-122 Metro 184 4121-132 ITC Berkeley Old Style 130 7267-513 Oliver 185 4133-122 Elante 131 7261-113 ITC Franklin Gothic 186 4133-162 Electra 132 7361-113 Helvetica 187 4145-212 Comenius Antiqua 133 7361-132 Folio 188 4223-112 Claridge 134 7361-212 Univers 189 4233-112 ITCClearface 135 7361-213 Heldustry 190 4233-112 Quadriga-Antiqua 136 7361-412 Eurostile 191 4245-122 Basilia 137 7364-112 Akzidenz Grotesk 192 4245-122 Walbaum 138 7365-523 Hobo 193 4253-122 Bauer Bodoni 139 7367-113 ITC Mixage 194 4255-113 ITCFenice 140 7461-112 Univers Extended 195

248 CLASSIFICATION INDEX 63

7461-113 Helvetica Extended 196 7513-113 Clearface Gothic 197 7523-132 Globe Gothic 198 7561-113 ITC Avant Garde Gothic Condensed 199 7561-113 Franklin Gothic Condensed 200 7561-113 Helvetica Condensed 201 7561-113 Trade Gothic Condensed 202 7561-312 Akzidenz Grotesk Condensed 203 7561-312 Univers Condensed 204 7561-313 Alpin Gothic 205 7561-313 Helvetica Compressed 206 7561-413 Compacta 207 7561-614 ITC Machine 208 7567-112 Futura Condensed 209 7567-112 Gill Sans Condensed 210 7567-142 Spartan Condensed 211 8164-112 Symphony 212 8164-112 Syntax 213 8167-123 ITC Kabel 214 8664-113 Orator 215 9111-123 Albertus 216 9111-123 Romic 217 9111-133 Friz Quadrata 218 9123-122 Optima 219 9167-113 ITC Symbol 220 9167-123 ITC Elan 221 9167-123 ITC Serif Gothic 222 9168-212 ITC Goudy Sans 223 9223-132 Souvenir Gothic 224 9223-213 Poppl-Laudatio 225 9267-113 Flange 226 9267-113 Icone 227 9267-213 ITC Quorum 228 9269-532 Shannon 229 9313-113 Bluejack 230 9325,-413 Modula 231 9364-113 American Gothic 232 9364-114 Copperplate Gothic 233 9467-113 ITC Newtext 234 9523-213 Poppl-Laudatio Condensed 235 0164-183 Wexford 236 0167-122 Harry 237 0167-153 ITC Bauhaus 238 0268-143 ITC Benguiat Gothic 239 0364-113 Helios Rounded 240 64

Appendix C

SAMPLES OF PANOSE SPECIMEN PAGES 65

USING THE SPECIMEN PAGES

At the top of each specimen page are the font's listing to the left, to visualize the appearance of the name and classification number. Below the font other thirty-six. Beneath each of the display sam name is an alphabet printed in a medium weight ples is a name label noting the full name of the (if available). The entire alphabet has been printed preceding font. vertically down the page so that the page can be If an unusual stylistic change occurs because of folded back for easy comparison with a font on a change in weight, it will either be displayed on the different page. Red circles are drawn around iden specimen page or a separate classification number tifying characteristics of each letter to indicate will be listed and its specimen page noted above what makes this style different from others that the font information chart. If a type style has com might share this font's classification number. panion alphabets, such as condensed or expanded, Iii order to demonstrate the appearance of the these variations will usually have separate classifi type style in different weights, the uppercase letters cation numbers and their specimen pages can be P, A, N, O, S, and and the lowercase letters a, h, found in the cross-reference chart. If there are not e, g, k, m, o, q, s, and t are printed in all available sufficient weight variations in a companion alpha weight variations beneath the classification num bet to warrant a separate specimen page, the vari ber. These sixteen letters were selected because ations that do exist will be shown on the same page they represent the various basic letterforms, mak that displays the roman (noncondensed, nonex- ing it easy, in conjunction with the full alphabet panded) style of the font. 66

USING THE REFERENCE INFORMATION

Below each specimen display is a cross-reference able. Although most hot-type casting methods are chart. When you are selecting a type style, there is no longer used, many unique hot styles are in often one characteristic that you wish you could cluded in this book because digital fonts, which are alter. You could spend many hours paging through entering the market at an astonishing rate, are an ordinary specimen book in search of that "just often remakes of popular hot fonts of the past.

right" style, or you could use the PANOSE cross- There is no reason to assume that a popular font reference chart, which directs you to similar type will be discontinued because of the changes in ty styles if they exist. This chart, at the bottom of pographic technology. Use the information in this each page, offers digit-by-digit variations of the list as a minimum listing of the different media current classification number and lists the corre currently available. The media included are: hot foundry- sponding display page numbers of these slightly (including both and line-casting meth altered styles. Often the pages listed on the cross- ods); photo (any method where a film negative reference chart merely reflect an adequate alterna font is projected onto light-sensitive paper); dry tive that is close to the specifications you desire. If, transfer (adhesive sheets of characters to be on the other hand, the page listed on the chart is pressed onto paste-up boards); digital (fonts held circled in red, this alternate is an identical digit in a computer as geometric formulas and imaged match to your current classification number, with onto light-sensitive paper with a directed laser the exception of the single digit for which you have beam); and bit-mapped (fonts composed of dots specified a change. on a grid; these fonts are used on both computer In addition to the cross-reference chart, other monitors and high-resolution laser printers). Con information is given about the font in a box above cerning the two computer media, digital and bit and to the right of the chart. The first information mapped, two additional distinctions are noted. in the box is the name of the vendor who supplied PostScript is a new page description language

the specimen for the font display. Since some font which is capable of very sophisticated type manip names are unprotected, many type styles with the ulation. Many popular fonts are being converted same names are slightly altered by different sup into this format. These fonts are distinguished by whose media Bitstream Inc. pliers to suit their needs. Knowing exactly a I'S in the listing. Also, has font you are looking at can be very handy. digitizing fonts to create a library of high-quality The second line of information contains the outline font descriptions. These outlines can be characters for a wide type's proposed use as display or text. Often both converted into bit-mapped of digital and video are plausible uses; the most appropriate use is variety laser printer, typesetter, listed first. Usually this purpose is self-evident. applications. These fonts are listed with a BT. will continue to be. reworked Similar names are then provided, listing type More popular faces in as advances. faces that are visually related and can be found technology recommends a different sourcebooks. The listed names are usually The similar fonts list generally appearance yet less conventional and more ornate. PANOSE does few fonts which are different in styles. similar in to the displayed font. These are not attempt to classify decorative type feeling information often popular fonts with classification num Below the similar-names listing is less that either were on the media for which this font is currently avail bers similar to the displayed font, 67

not able to be included in this book or are located on a nonadjacent page.

Also provided is a list of alternate names. These names often reflect fonts that differ only slightly from the shown font. In these cases an abbrevia tion is given to specify the supplier who uses this alternate name. A listing of the abbreviations for various suppliers precedes the bibliography. Inas much as a single font name will vary, from supplier to supplier, so too will the alternate names listed in this book. These names are not intended to sig nify identical styles but rather related ones that could not be displayed. 68

3121-112- Classification Font Name - Palatino Number ABCBEFG ABCDEFGH1JKLMN Full Alphabet, OPQRSTUVWXYZ Medium abcdefghijklm \ Variations of Weight, -HIJKLMN nopqrstuvwxyz Weight and Set in 48 point ^234567890. &!? , 7 Set When Possible 1^QRS:T Proportion, ABCDEFGHIJKLM in 24 point when NOPQRSTUVWXYZ possible abcdefghijklm s nopqrstuvwxyz Identification abcdefghij 1234567890.,;:"&!? Additional Marks ^miippqr PANOSE abegmoqs Information hMHWUMMlMM s^ivwxyz Including: The Supplier of This

1234567890 cwifiBWc Mao aim Twrnnw Sample, Name Notes, U*aye Ttrt,DtopllT .in Proposed Similar l-unls mm Usage, the Wim HawM. Ca Describing &!?$ ulablc Media Available Media, True Font Name, wn. owai. in. mi. in. n Alternate Name* Similar Names, lift chiiwmmc mim. knHkM. una. Dim). Weight, Style, PMcMM, HUMdUUI, EtaffMOIC), and Alternate IIM|IW. PmUeQUMtl and Point Size Names

Cross Reference Chart

Page Number 69

Baskerville 1133-122 ABCDEFGHIJKLMNO ABQDEFG PQRSTUVWXYZ H1JKLMN abcdefghijklmnopqrstu vwxyz 1234567890.,;:s&!?$ OPQRSTU Baskerville Hallo 24 point ABCDEFGHIJKLMN VWXYZ ab OPQRSTUVWXYZ abcdefghijklmnopqrst

cdelghijklmn uvwxyz 1234567890.,;:'^!?$ Baskerville Bold 24 point

opqrstuvwxy PANOSE abegmoqst 1234567890 Baskerville Bold Italic 24 point

Supplier Alphatype CHS Digital Typesetter Usage Text, Display Baskerville 48 point Similar Fonts Baskerville No.2, Baskerville 353 Available Media Digital, Photo, Hot, PT, Bit Alternate Names _ fractional 401(BS)7BK(ltek), Baskerville ll(CC), Beaumont(Wang)

1 2 3 4 5 6 7 8 9 0 serif X ^95J 4131 X 161 X X X X proportion (M) 56 X X X ~v contrast 2 ch> 35 X X arm style 30 X X X X X X X

fc -.. form X X X X X .""-=.-'%

V' midline X X X X X X X

x height X X X

31 70

ITC Serif Gothic 9167-123 ADCDEFGHIJKLMNO ABCDEFd PQRSTUVWXYZ abcdefghijklmnopq HIJKLMNO rstuvwxyz 1204567d90.,;vfc!?$ PQRSTUV ITC Serif Gothic Heavy 24 point PANOSE obegmoqsr WXYZcfbc ITC Strlf Gothic light 24 point PANOSE abegmoqst 4efghijkl ITC Swif Gothic Bold 24 point PANOSE abegmoqst mnopqrsf ITC Sortf Gothic Extra Bold 24 point PANOSE abegmoqst

ITC Sortf Gothic Black 24 point uvwxyz

1 234567 Supplier Alphatype CRS Digital Typesetter Usage Display Similar Fonts Sleek Gothic Available Media 890.,;w&!?$ Digital, DT, Photo, Bit ITC Snrtf Gothic 48 point Alternate Names Une BothicOrVang), Copperplate 701 (BS)

1 2 3 4 5 6 7 8 9 0 serif X 69 96 X X X 171 214 237

f- proportion 226 X 234 X X *mA r-

contrast X X X X X 4iuG sj^fctaEtHr- SaSft ku

rT arm style X X X X X X 223 X

form X X X X X :>>i, sts&fe bA fe&g* War midline 220 X X X X X X ^#is 9fetV >V"fcs. **"' x height X X X

222 71

Appendix D

TYPE-ID PROFILE GENERATOR 72

PROFILE GENERATOR KEY

The first part of the underscored first line indicates from which sequence the questions comes. There will The question eventually be a sequence for each number appears letter, figure and punctuation mark. after the sequence label.

4V -7 SERIF GENERAL QUESTION #1

The lowercase ascenders are generally the capital letters (including serifs).

/ (A) TK Each illustration corresponds to a written answer (B) below. The key points of each Th illustration are indicated with arrows. The answers are / in bold type. See illustrations for (C) clarification of terms used. TR \ A. taller than B. the same height as C. shorter than 73

SERIF GENERAL QUESTION #1

The lowercase ascenders are generally the capitals letters (including serifs).

(A)

(B)

(Q

A. taller than B. the same height as C. shorter than 74

SERIF GENERAL QUESTION #2

Are the characters in your typeface composed of strokes of equal thickness (excluding serifs) or do the strokes vary in thickness?

(A)

(B)

A. strokes of equal thickness B. strokes of varying thickness 75

SERIF GENERAL QUESTION #3

"e" The crossbar of the lowercase is:

(A)

(B)

A. horizontal B. angled 76

SERIF GENERAL QUESTION #4

The caps E, F and P are the "round" caps such as C, D, G, O, and Q.

(A) CEDFOP

(B) CEDFOP

A. noticably narrower than B. approximately the same width as 77

SERIF GENERAL QUESTION #5

What is the angle of stress?

stress" The 'angle of can be determined by drawing a line through the thinnest portions of such round characters as (A) the lowercase o, e and c (as shown in illustrations 1-3 at left). Characters which are composed of strokes of equal

thickness have no angle of stress (illustration 4).

(B)

(C)

(D)

A. vertical B. acute C. grave D. not applicable (eg. all strokes same thickness) 78

CAP "A" QUESTION #1

"top" "apex" The or of the Cap A is:

(A)

(B)

A

_ / ,0,7V/ AAA

A. pointed or rounded B. full-seriffed C. half-seriffed D. other (see examples) 79

CAP "A" QUESTION #2

"top" "apex" At the or of the cap A does one stem extend beyond the other?

(A) A - A (XX

A. yes B. no 80

CAP "A" QUESTION #3

The bar is:

(A)

Note: Determine the position of the bar by

space" its relationship to the "white inside the A (see illustration below). If distance a is approximately equal to distance b then the (d) / \ bar is centered. If a is less than b then the bar is high. If a is greater than b the bar is A low. If the orientation of the bar is anything other than horizontal, choose choice 4, other.

(C)

(D)

A. centered B. high C. low D. other (including all non-horizontal bars) 81

CAP "A" QUESTION #4

Left stem (left edge):

A. straight B. concave C. convex 82

CAP "A" QUESTION #5

Right stem (right edge):

(A)

(B)

(C)

A. straight B. concave C. convex 83

CAP "A" QUESTION #6

Does the left foot serif have "brackets"?

(A) Brackets are structures which, when present,

"soften" off" or "round

the angle between a serif

and its stem.

(B)

No brackets.

A. Yes B. No 84

CAP "A" QUESTION #7

Does the right foot serif have "brackets"?

when ( A\ \ ^L ^s Brackets are structures which, Vv \ bA. ^ "soften" "round-off" present, or the angle between a serif and its stem.

(B) No brackets.

A. Yes B. No 85

CAP"C" QUESTION #1

How many serifs are there?

(A)

(B)

(C)

A. one B. two C. none 86

CAP"C" QUESTION #2

Does the top serif extend above its arm?

(A)

(B)

A. yes B. no 87

CAP"C" QUESTION #3

Does the bottom serif extend below its arm?

(A)

(B)

A. yes B. no 88

CAP"E" QUESTION #1

Top arm is middle arm (including serifs).

(A)

(B)

A. longer than B. same length as 89

CAP"E" QUESTION #2

Middle arm is bottom arm (including serifs).

(A)

(B)

A. same length as B. shorter than 90

CAP"E" QUESTION #3

Bottom arm is top arm (including serifs).

(A)

(B)

A. longer than B. the same length as 91

CAP "E" QUESTION #4

Top-arm serif orientation (right edge):

(A)

(B)

(C)

A. vertical B. grave C. acute 92

CAP"E" QUESTION #5

Does any part of the top-arm serif extend above the cap line?

(A)

(B)

A. yes B. no 93

CAP"E" QUESTION #6

Bottom-arm serif orientation (right edge):

(A) E.

(B)

&/

(C)

(D)

A. vertical B. acute C. grave D. curved or hooked arm with no serif 94

CAP"E" QUESTION #7

Does any part of the bottom-arm serif extend below the baseline?

A. yes B. no 95

CAP"G" QUESTION #1

"G" "spur" Does the cap have a or "nib"?

v-*--. (A, Q_ Note: Spurs and n/6s are

protrusions.

spur

(B) G-G

A. yes B. no 96

CAP"G" QUESTION #2

Lower serif description:

(A)

(B)

A. full serif B. half-serif (left) 97

CAP"G" QUESTION #3

Lower arm (right edge) description:

(A)

(B)

A. curved B. straight 98

CAPT QUESTION #1

"hook" "J" Does the of the cap extend below the baseline?

(A)

(B)

A. yes B. no 99

CAP"L" QUESTION #1

"arm" Is the bottom edge of the of "L" the cap (excluding serifs) straight?

(A)

\

(B)

\

A. straight B. other (eg. concave) 100

CAP"L" QUESTION #2

Arm serif orientation (right edge):

(A)

(B)

(C)

(D)

A. vertical B. acute C. grave D. rounded or hooked 101

CAP"M" QUESTION #1

The left apex has:

\

(A)

\

(B)

\

(Q

A. full serif B. half serif C. no serif 102

CAP"M" QUESTION #2

The right apex has: /

(A)

/

(B)

/

(C)

A. full serif B. half serif C. no serif 103

CAP"M" QUESTION #3

"vertex" Does the reach the baseline?

(A) \J

(B)

A. yes B. no 104

CAP"M" QUESTION #4

Vertex is:

(A) or

(B) 1M or

A. pointed or rounded B. other (see examples) 105

CAP"M" QUESTION #5

Left stem (left edge) orientation (excluding) serifs:

(A)

(B)

(C)

(D)

A. vertical B. acute C. concave D. convex 106

CAP"M" QUESTION #6

Right stem (right edge) orientation excluding serifs:

(A)

(B)

(C)

(D)

A. vertical B. acute C. concave D. convex 107

CAP"P" QUESTION #1

Is the bowl completely enclosed?

(A)

\

(B)

A. yes B. no 108

CAP"Q" QUESTION #1

Does any part of the tail reach inside the bowl?

(A)

(B)

A. yes B. no 109

CAP"R" QUESTION #1

Does the tail extend below the baseline?

(A)

\

(B)

\

A. yes B. no 110

CAP "R" QUESTION #2

Is the bowl completely enclosed?

(A)

(B)

A. yes B. no 111

CAP"R" QUESTION #3

Tail orientation (not including serif):

(A) R (B) R (C) R

A. straight B. concave C. convex 112

CAP"T" QUESTION #1

Left arm (left edge) serif orientation:

i (A) T

(B) T

\ (C) T

A. vertical B. acute C. grave 113

CAP"T" QUESTION #2

Right arm (right edge) serif orientation:

(A)

(B)

(C) f

A. vertical B. acute C. grave 114

CAP"T" QUESTION #3

How many serifs extend above the arm?

\ / (A)

or

\ /

"""" (B)

(C)

A. one B. two C. none 115

CAP"W" QUESTION #1

Is the left stem (excluding serifs) basically straight?

(A)

(B)

A. yes B. no (eg. rounded stem) 116

CAP"W" QUESTION #2

Is the right stem (excluding serifs) basically straight?

(A)

(B)

A. yes B. no (or other) 117

CAP"W" QUESTION #3

Which of the following best describes the "branches" arrangement of the (ie. the two inner-most strokes) of the cap W?

JOINED: Both branches meet and end at the same

(A) point or continue on jointly. After the

junction point one branch cannot be

distinguished from the other.

STEPPED: (B) After the junction

point, one branch

continues, the other ends. w

CROSSED: both _, / (v_-/ / branches continue on after the junction

point.

A. joined B. stepped C. crossed 118

CAP "W" QUESTION #4

How many top serifs are there?

(A)

(B)

(Q w-w-w

(D) Note: The illustrations

are examples only.

Other configurations may exist. (E) w

A. one B. two C. three D. four E. other (eg. no serifs) 119

CAP"W" QUESTION #5

"shared" Are any of the top serifs (ie. touch more than one branch or stem)?

shared serif shared serjf / / (A) WW

The serifs in this illustration are all

"individual" i i r considered to be serifs. The center serif may, in some instances, touch more than one branch, but as (f3) \ g\ f long as these branches are either "joined" "stepped" or and not

"crossed" the center serif is considered

to be an individual serif. (See CAP W QUESTION # 3 for definitions of

"joined," "stepped," W and "crossed.") (C) w

A. yes B. no C. not applicable (eg. no serifs) 120

CAP"W" QUESTION #6

The vertexes (bottom junctions) are:

(A) or w t t t t

(B) w I I

w' -'i i

w w? ? II

A. pointed or rounded B. other (see examples) 121

LOWERCASE "a" QUESTION #1

Number of stories:

(A) a

(B)

A. one B. two 122

LOWERCASE "a" QUESTION #2

Tail:

(A) J* or

(B)

..^i.

(C)

(D)

A. curved B. flat half-serif C. angled half-serif D. other (eg. no serif) 123

LOWERCASE "a" QUESTION #3

Is the bowl completely enclosed?

(A)

/

(B)

A. yes B. no 124

LOWERCASE "a" QUESTION #4

41Shoulder" terminal:

(A)

(C)

(D) or V^ (E)

(F)

(C) a

A. ball or tear-drop B. hooked and pointed C. blunt, flat D. flared or half-seriffed E. rounded F. other ornament (eg. flag) G. not applicable (eg. no shoulder) 125

LOWERCASE "e" QUESTION #1

The bowl/bar junction is:

-e -e

(C)

(D)

A. pointed B. angled C. spurred D. rounded 126

LOWERCASE "f" QUESTION #1

Shoulder terminal:

r- (A)

(B)

(C)

(D) fpyffi

(E)

r- (F)

A. ball or tear-drop B. hooked and pointed C. blunt D. flared or half-seriffed E. rounded F. other ornament (eg. flag) 127

LOWERCASE "f" QUESTION #2

The crossbar appears to be:

(A) H (D)

+ (B) (E)

(C) -4 (F)

A. centered on the stem B. more right C. more left D. right only E. left only F. not applicable (eg. no crossbar) 128

LOWERCASE "g" GENERAL QUESTION

Number of stories:

(A)

(B)

A. one

(Please turn to pages 71-73 and

answer the questions for the

"g" ONE-STORY LOWERCASE

only. Do not answer those

questions labeled TWO-STORY

"g" LOWERCASE

B. two

(Please turn to pages 74-79 and

answer the questions for the "g" TWO-STORY LOWERCASE

only. Do not answer those

questions labeled ONE-STORY LOWERCASE "g". 129

ONE-STORY LOWERCASE "g" QUESTION #1

Does the top of the stem have a serif?

/ ^"^ (A)

/

(B)

A. yes B. no 130

ONE-STORY LOWERCASE "g" QUESTION #2

Is the bowl completely enclosed?

/

(A)

(B)

A. yes B. no 131

ONE-STORY LOWERCASE "g" QUESTION #3

Does the end of the tail have an ornament?

(A)

The tail is considered to have an ornament if it has a serif, ball, pear, hook or any other pronounced flourish at its end.

(B)

A. yes B. no 132

TWO-STORY LOWERCASE "g" QUESTION #1

"ear" The end of the is:

An ear is an ornament which protrudes from the upper bowl.

The ear is (A) unfinished (D) if it has the

same thick ness or nearly the same thickness for

its entire length.

(B) (E)

(C) (F)

GENERAL NOTE: The ball- or pear-shaped A. angle of the ear need not B. pointed (or tapered) be the same as indicated flared in the above examples. C. Vertical, horizontal or D. unfinished other orientations are possible for each of the other (eg. E. flag) choices. F. not applicable (eg. no ear) 133

TWO-STORY LOWERCASE "g" QUESTION #2

Does the ear curve?

(A)

(B)

(C) a

(D)

A. yes, clockwise B. yes, counter-clockwise C. no (ear is straight) D. other (eg. no ear) 134

TWO-STORY LOWERCASE "g" QUESTION #3

Is the upper bowl completely enclosed?

/

(A)

/

(B)

A. yes B. no 135

TWO-STORY LOWERCASE "g" QUESTION #4

Is the bottom bowl completely enclosed?

(A)

/

(B)

/

A. yes B. no 136

TWO-STORY LOWERCASE "g" QUESTION #5

Does the bottom bowl have an ornament?

(A) The bottom bowl is

considered to have an

ornament if it has a serif, ball, pear, hook or any other pronounced flourish at its end.

(B)

(C)

A. yes B. no C. not applicable (bottom bowl enclosed) 137

LOWERCASE "r" QUESTION #1

Does the end of the arm have a ball- or pear-shaped terminal?

(A)

(B)

A. yes B. no 138

LOWERCASE "r" QUESTION #2

Top stem serif orientation (top edge)

a.

Note: if point a is lower than point b then the top serif (A) orientation is acute. If point a is above point b then the top serif orientation is grave. If points a and b are in the same plane, the top serif orientation is horizontal. This is true whether the top serif connecting points a and b is straight or curved.

(B) r

\ / (C)

A. horizontal B. acute C. grave 139

FIGURE "1" QUESTION #1

Head serif orientation (top edge):

^ ^ Note: if point a is lower than point b then the head serif

orientation is acute. If point a is above point b then the

_ head serif orientation is grave. If points a and b are in \r\) | the same plane, the head serif orientation is horizontal. This is true whether the head serif connecting points a 1 and b is straight or curved.

A. horizontal B. acute C. grave 140

FIGURE "2" QUESTION #1

Shoulder terminal:

(A>cr^ (B)-

(O-

(D)- -y^

(E)-

(F)

A. ball or pear B. point C. blunt, flat D. flared or seriffed E. rounded F. other (eg. flag) 141

FIGURE "2" QUESTION #2

Is the bottom edge of the tail straight (excluding serifs or ornaments)?

(A)

(B)

A. yes B. no (eg. curved) 142

"2" FIGURE QUESTION #3

The end of the tail is:

(A)

(B)

(C)

(D)

(E)

(F)

A. pointed B. flat or blunt C. half-serifed (up) D. half-serifed (down) E. full-seriffed F. other (eg. hooked) 143

FIGURE "3" QUESTION #1

Total number of strokes (excluding serifs)

(A)

1_V (B)

A. two B. three 144

FIGURE "3" QUESTION #2

Does the figure 3 extend below the baseline?

(A)

(B)

\

A. yes B. no 145

FIGURE "4" QUESTION #1

Top apex:

\

(A)

\

(B)

A. open B. closed 146

FIGURE "4" QUESTION #2

Does the top apex have a serif?

(A)

(B)

"open" NOTE: If the apex is

then choose choice 3 (not

applicable) even if there is a

serif structure of some sort

present. (C)

A. yes B. no C. not applicable (eg. open apex) 147

FIGURE "4" QUESTION #3

Crossbar terminal (for horizontal crossbars only):

(A) 4 (B) 4 (C) 4 (D) 4

orientation of ,c. - - NOTE: If the the ^^ *^^ \t) ^^T^ crossbar is anything other than horizontal choose answer 5 (other) or not. 4 whether the crossbar has a serif

A. half serif up B. half serif down C. full serif D. no serif E. other (see note) 148

PERIOD (.) QUESTION #1

The shape of the period is:

(A) or

(B)

(C) +

(D)

A. dot, circle, oval B. square C. diamond D. other 149

Appendix E

TYPEFACE SPECIMENS 150

TYPEFACES

ADMINSTER BOOK (C) CALLIARD ROMAN (ITC) (V) AMERICAN TYPEWRITER LIGHT (L) GARAMOND (V) ANDOVER (V) GARAMOND NO. 3 (L) ARTCRAFT LIGHT (C) GARTH GRAPHIC (C) AURORA TTS 8 SET (L) GOUDY (V) BASKERVILLE (V) GOUDY OLDSTYLE (L) BASKERVILLE BOLD (L) JANSON (C) BASKERVILLE II (C) MELIOR (L) BEMBO (L) MEMPHIS (L) BODONI (L) MODERN (V) BODONI BOOK (V) NEWS NO. 9 (C) BOOKMAN (L) PALATINO (L) CALEDONIA (L) PLANTIN (C) CARTIER (C) RALEIGH (A) CASLON # 3 (C) ROCKWELL (V) CASLON #224 MEDIUM (ITC) (V) SABON (C) CENTURY OLDSTYLE (V) SCHNEIDLER(C) CENTURY SCHOOLBOOK (L) SERIF GOTHIC LIGHT (ITC) (V) CENTURY TEXTBOOK (A) SOUVENIR (V) CHELTENHAM BOLD (C) STYMIE LIGHT (C) CLARENDON BOOK (C) TIFFANY (ITC) (V) CLEARFACE BOLD (A) TIMES BOLD (CC) (C) COLONIAL (V) TIMES ROMAN (V) CONGRESS MEDIUM (C) TRUMP MEDIEVAL (C) COPPERPLATE LIGHT 4 (C) WINDSOR LIGHT (C)

KEY

Atex Fonts (A) Compugraphic Fonts (c) Linotype Fonts (L) Varityper Fonts (V) Century Textbook Adminster Book American Typewriter Light Andover Clearface Bold Artcraft Light Aurora TTS 8 set Baskerville Raleigh Baskerville II Baskerville Bold Bodoni Book Medium Cartier Bembo Caslon #224 (ITC) Caslon #3 Bodoni Century Oldstyle Cheltenham Bold Bookman Colonial Roman Clarendon Book Caledonia Galliard (ITC) Congress Medium Century Schoolbook Garamond Copperplate Light 4 Garamond No. 3 Goudy Garth Graphic Goudy Oldstyle Modern Rockwell Janson Melior Serif Gothic Light News No. 9 Memphis (ITC) Souvenir Plantin Palatmo Sabon Tiffany (ITC) Times Roman Schneidler Stymie Light Times Bold (CG) Trump Medieval Windsor Light 151

ADMINSTER BOOK (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 152

AMERICAN TYPEWRITER LIGHT (48-point)

ABGDEFGHI

QJ IlUlVll OPQR vwxy:

cdefgfriiklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukcia nor 153

ANDOVER (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 154

ARTCRAFT LIGHT (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcde%hij klm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 155

AURORA TTS 8 SET (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 156

BASKERVILLE (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 157

BASKERVILLE BOLD (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 158

BASKERVILLE II (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 159

BEMBO (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 160

BODONI (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 161

BODONI BOOK (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 162

BOOKMAN (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 163

CALEDONIA (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 164

CARTIER (48-point)

ABCDEFGHI

JKLMNOPQR

STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 165

CASLON #3 (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 166

CASLON #224 MEDIUM (ITC) (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 167

CENTURY OLDSTYLE (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 168

CENTURY SCHOOLBOOK (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 169

CENTURY TEXTBOOK (48-pointi

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxy

1234567890?!,.

Eht Ukciq Wbnor 170

CHELTENHAM BOLD (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 171

CLARENDON BOOK (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 172

CLEARFACE BOLD (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxy

1234567890?!, .

Eht Ukciq Wbnor 173

COLONIAL (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1 234567890?!,.

Eht Ukciq Wbno 174

CONGRESS MEDIUM (48-point)

ABCDEFGHI

JKLMNOPQR

STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 175

COPPERPLATE LIGHT 4 (48-point)

ABCDEFGHI

JKLMNOPQR

STUVWXYZ

ABCDEFGHIJKLM

NOPQRSTUVWXYZ

1234567890?!,.

Eht Ukciq Wbnor 176

GALLIARD ROMAN (ITC) (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 177

GARAMOND (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 178

GARAMOND NO. 3 (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 179

GARTH GRAPHIC (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 180

GOUDY (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 181

GOUDY OLDSTYLE (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 182

JANSON (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 183

MELIOR (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 184

MEMPHIS (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 185

MODERN (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghij klmnopqrs tuvwxyz

1234567890?!,

Eht Ukciq Wb 186

NEWS NO. 9 (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 187

PALATINO (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 188

PLANTIN (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 189

RALEIGH (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxy

1234567890?!,.

Eht Ukciq Wbnor 190

ROCKWELL (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 191

SABON (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 192

SCHNEIDLER (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 193

SERIF GOTHIC LIGHT (ITC) (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Ehr Ukciq Wbnor 194

SOUVENIR (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 195

STYMIE LIGHT (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 196

TIFFANY (ITC) (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm

nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 197

CG TIMES BOLD (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 198

TIMES ROMAN (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 199

TRUMP MEDIEVAL (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ

abcdefghijklm nopqrstuvwxyz

1234567890?!,.

Eht Ukciq Wbnor 200

WINDSOR LIGHT (48-point)

ABCDEFGHI JKLMNOPQR STUVWXYZ abcdefghijklm nopqrstuvwxyz

1 %

Eht Ukciq Wbnor 201

Appendix F

DATA 202

TABLE 1 KEY

Collected Data

UNIDENTIFIED TYPEFACE NAME: SEQUENCE: Indicates which Names of the 50 typefaces with series of questions the user which subjects tested the TYPE-ID answered. There are 20 system are entered in this column. sequences in all. The (Profiles of the same 50 typefaces questions for each sequence are stored in the database. can be found in Appendix D.

S E Q U N C E

UNIDENTIFIED TYPfeFACE CAP A CAPC CAP J CAP L name

Search Match Search Match Search Match Search Match Search Match Yield Found? Yield Found? Yield Found? Yield Found? Yield Found?

Vdminster Book. 45J (Yes) Yes 32 Yes 29 Yes 37 Yes Amercn. Typwrtr. Lt. Yes\ 32 Yes 12 Yes 23 Yes

SEARCH YIELD: Total number of MATCH FOUND?: Is the name of "list" typefaces in the generated by the user's unidentified typeface "yes," the computer which have the same included in the list? If so, if "no" search argument as the user's not, is entered in this column. "45" unidentified typeface. Here, Here, Adminster Book was included on the list. typefaces from the database of 50, among the 45 typefaces satisfied the search argument. 203

TABLE 1 - Collected Data

S E Q U E N C E UNIDENTIFIED TYPEFACE CAP A CAPC CAPG CAP J CAPL NAME

Search Match Search Match Search Match Search Match Search Match Yield Found? Yield Found? Yield Found? Yield Found? Yield Found?

Adminster Book 43 yes 4 yes 32 yes 29 yes 37 yes

American Typewriter 26 yes 22 yes 32 yes 18 yes 23 yes

Andover 18 yes 8 yes 24 yes 39 yes 37 yes

Artcraft Light 29 yes 17 yes 44 yes 39 yes 50 yes

Aurora 18 yes 22 yes 30 yes 50 yes 50 yes

Baskerville 20 yes 19 yes 22 yes 39 yes 23 yes

Baskerville Bold 36 yes 19 yes 40 yes 39 yes 23 yes

Baskerville II 29 yes 19 yes 40 yes 39 yes 50 yes

Bembo 43 yes 19 yes 22 yes 39 yes 37 yes

Bodoni 2 yes 19 yes 5 yes 39 yes 50 yes

Bodoni Book 2 yes 19 yes 5 yes 39 yes 23 yes

Bookman 16 yes 22 yes 30 yes 39 yes 37 yes

Caledonia 26 yes 21 yes 40 yes 39 yes 50 yes

Cartier 13 yes 8 yes 40 yes 39 yes 37 yes

Caslon #3 11 yes 19 yes 22 yes 50 yes 37 yes

Caslon #224 16 yes 19 yes 30 yes 18 yes 37 yes

Century Oldstyle 16 yes 22 yes 22 yes 39 yes 37 yes

yes yes 50 yes Century Schoolbook 27 yes 22 yes 30 18

50 yes 50 yes Century Textbook 27 yes 22 yes 30 yes

yes 50 yes Cheltenham Bold 43 yes 22 yes 4 yes 50

yes 50 yes Clarendon Book 17 yes 22 yes 32 yes 50

50 yes 37 yes Clearface Bold 15 yes 22 yes 45 yes

18 yes 37 yes Colonial 43 yes 10 yes 22 yes

39 yes 50 yes Congress Medium 43 yes 26 yes 45 yes

18 yes 23 yes Copperplate Gothic 14 yes 19 yes 34 yes 204

TABLE 1 (Continued)

S E Q U E N C E UNIDENTIFIED TYPEFACE CAP A CAP C CAPG CAP J CAP L NAME

Search Match Search Match Search Match Search Match Search Match Yield Found? Yield Found? Yield Found? Yield Found? Yield Found?

* Galliard Roman no 10 yes 30 yes 39 yes 37 yes

* Garamond no 10 yes 22 yes 39 yes 37 yes

Garamond No. 3 27 yes 25 yes 5 yes 39 yes 37 yes

Garth Graphic 25 yes 26 yes 6 yes 39 yes 37 yes

Goudy 20 yes 10 yes 22 yes 39 yes 14 yes

Goudy Oldstyle 26 yes 10 yes 6 yes 39 yes 14 yes

Janson 22 yes 21 yes 40 yes 39 yes 37 yes

Melior 18 yes 25 yes 40 yes 39 yes 50 yes

Memphis 5 yes 22 yes 30 yes 39 yes 23 yes

Modern 2 yes 22 yes 30 yes 18 yes 23 yes

News No. 9 26 yes 22 yes 30 yes 50 yes 50 yes

Palatino 36 yes 26 yes 22 yes 39 yee 37 yes

Plantin 15 yes 25 yes 40 yes 39 yee 37 yes

Raleigh 16 yes 26 yes 40 yes 39 yee 23 yes

Rockwell 5 yes 22 yes 1 yes 39 yee 23 yes

Sabon 13 yes 25 yes 32 yes 39 yee 37 yes

37 yes Schneidler 16 yes 25 yes 40 yes 39 yes

* 23 yes Serif Gothic Light 15 yes 19 yes no 18 yes

yes 18 yes 37 yes Souvenir 2 yes 8 yes 4

39 yes 23 yes Stymie Light 5 yes 22 yes 30 yes

18 yes 37 yes Tiffany 13 yes 22 yes 30 yes

yes 39 yes 37 yes Times Bold 20 yes 22 yes 22

18 yes 37 yes Times Roman 20 yes 22 yes 22 yes

30 yes 39 yes 37 yes Trump Medieval 24 yes 25 yes 30 yes 18 yes 37 yes Windsor Light 16 yes 22 yes 205

TABLE 1 (Continued)

S E Q U E N C E UNIDENTIFIED TYPEFACE CAP M CAP P CAP Q CAP R CAPT NAME

Search Match Search Match Search Match Search Match Search Match Yield Found? Yield Found? Yield Found? Yield Found? Yield Found?

Adminster Book 36 yes 11 yes 35 yes 7 yes 12 yes

American Typewriter 28 yes 38 yes 18 yes 19 yes 21 yes

Andover 28 yes 11 yes 35 yes 2 yes 22 yes

* Artcraft Light 6 yes 11 yes 18 yes no 14 yes

Aurora 14 yes 38 yes 18 yes 19 yes 25 yes

Baskerville 28 yes 38 yes 35 yes 27 yes 31 yes

Baskerville Bold 28 yes 38 yes 35 yes 27 yes 21 yes

Baskerville II 28 yes 38 yes 35 yes 27 yes 25 yes

Bembo 13 yes 38 yes 35 yes 5 yes 12 yes

Bodoni 14 yes 38 yes 35 yes 19 yes 21 yes

Bodoni Book 28 yes 38 yes 35 yes 21 yes 21 yes

Bookman 14 yes 38 yes 35 yes 27 yes 12 yes

Caledonia 28 yes 38 yes 35 yes 19 yes 14 yes

Cartier 7 yes 38 yes 50 yes 27 yes 14 yes

Caslon #3 12 yes 38 yes 35 yes 19 yes 12 yes

Caslon #224 28 yes 38 yes 35 yes 19 yes 12 yes

25 yes Century Oldstyle 28 yes 38 yes 18 yes 27 yes

14 yes Century Schoolbook 14 yes 38 yes 18 yes 19 yes 25 yes Century Textbook 14 yes 38 yes 18 yes 19 yes 21 yes Cheltenham Bold 14 yes 38 yes 35 yes 19 yes

21 yes Clarendon Book 14 yes 38 yes 50 yes 19 yes

27 yes 7 yes Clearface Bold 5 yes 38 yes 18 yes

19 yes 25 yes Colonial 28 yes 38 yes 35 yes

27 yes 10 yes Congress Medium 7 yes 38 yes 35 yes

27 yes 3 yes Copperplate Gothic 2 yes 38 yes 18 yes 206

TABLE 1 (Continued)

S E Q U E N C E UNIDENTIFIED TYPEFACE CAPM CAP P CAP Q CAP R CAPT NAME

Search Match Search Match Search Match Search Match Search Match Yield Found? Yield Found? Yield Found? Yield Found? Yield Found?

Galliard Roman 13 yes 11 yes 35 yes 9 yes 12 yes

Garamond 13 yes 38 yes 35 yes 27 yes 5 yes

Garamond No. 3 3 yes 38 yes 35 yes 27 yes 5 yes

Garth Graphic 13 yes 38 yes 35 yes 19 yes 25 yes

Goudy 13 yes 11 yes 35 yes 27 yes 25 yes

Goudy Oldstyle 5 yes 11 yes 35 yes 27 yes 14 yes

Janson 13 yes 38 yes 35 yes 27 yes 12 yes

Melior 28 yes 38 yes 18 yes 27 yes 21 yes

Memphis 3 yes 38 yes 18 yes 19 yes 21 yes

Modern 28 yes 38 yes 18 yes 19 yes 21 yes

News No. 9 14 yes 38 yes 50 yes 19 yes 25 yes

Palatino 28 yes 11 yes 35 yes 2 yes 15 yes

Plantin 7 yes 11 yes 35 yes 27 yes 25 yes

Raleigh 28 yes 38 yes 35 yes 19 yes 21 yes

Rockwell * no 38 yes 35 yes 27 yes 21 yes

Sabon 13 yes 11 yes 35 yes 27 yes 7 yes

7 yes Schneidler 3 yes 11 yes 35 yes 5 yes

21 yes Serif Gothic Light 3 yes 38 yes 18 yes 1 yes

yes 25 yes Souvenir 8 yes 38 yes 18 yes 19

yes 19 yes 21 yes Stymie Light 6 yes 38 yes 35

27 yes 12 yes Tiffany 3 yes 38 yes 35 yes 27 yes 21 yes Times Bold 29 yes 38 yes 35 yes

27 yes 25 yes Times Roman 28 yes 38 yes 35 yes

18 yes 5 yes 25 yes Trump Medieval 13 yes 11 yes 27 yes 12 yes Windsor Light 2 yes 38 yes 18 yes 207

TABLE 1 (Continued)

SEQUENCE

UNIDENTIFIED TYPEFACE CAP W lea Ic e Icf leg NAME

Search Match Search Match Search Match Search Match Search Match Yield Found? Yield Found? Yield Found? Yield Found? Yield Found?

Adminster Book 12 yes 19 yes 47 yes 35 yes 5 yes

American Typewriter 6 yes 41 yes 47 yes 35 yes 20 yes

Andover 13 yes 12 yes 11 yes 21 yes 17 yes

Artcraft Light 12 yes 18 yes 47 yes 14 yes 7 yes

Aurora 9 yes 41 yes 47 yes 35 yes 20 yes

Baskerville 6 yes 41 yes 47 yes 35 yes 4 yes

Baskerville Bold 9 yes 41 yes 47 yes 35 yes 4 yes

Baskerville II 6 yes 41 yes 47 yes 35 yes 4 yes

Bembo 2 yes 19 yes 11 yes 21 yes 8 yes

Bodoni 2 yes 41 yes 47 yes 13 yes 20 yes

Bodoni Book 2 yes 41 yes 47 yes 13 yes 20 yes

Bookman 9 yes 41 yes 47 yes 35 yes 20 yes

Caledonia 12 yes 41 yes 47 yes 35 yes 5 yes

Cartier 1 yes 10 yes 11 yes 16 yes 8 yes

Caslon #3 12 yes 41 yes 47 yes 35 yes 20 yes

Caslon #224 12 yes 41 yes 47 yes 35 yes 20 yes

20 yes Century Oldstyle 12 yes 41 yes 47 yes 35 yes

yes 20 yes Century Schoolbook 2 yes 41 yes 47 yes 35

yes 20 yes Century Textbook 9 yes 41 yes 47 yes 35 4 no Cheltenham Bold 9 yes 41 yes 47 yes 35 yes

35 yes 20 yes Clarendon Book 9 yes 41 yes 47 yes

35 yes 20 yes Clearface Bold 2 yes 41 yes 47 yes

4 yes 6 yes Colonial 6 yes 12 yes 47 yes

16 yes 2 yes Congress Medium 9 yes 12 yes 47 yes

11 yes 50 yes Copperplate Gothic 9 yes 1 yes 50 yes 208

TABLE (Continued)

SEQUENCE

UNIDENTIFIED TYPEFACE CAP W Ic a Ic e Icf kg NAME

Search Match Search Match Search Match Search Match Search Match Yield Found? Yield Found? Yield Found? Yield Found? Yield Found?

Galliard Roman 12 yes 19 yes 47 yes 35 yes 17 yes

Garamond 7 yes 41 yes 47 yes 35 yes 17 yes

Garamond No. 3 7 yes 19 yes 47 yes 35 yes 17 yes

Garth Graphic 9 yes 12 yes 47 yes 21 yes 8 yes

Goudy 7 yes 19 yes 47 yes 21 yes 7 yes

Goudy Oldstyle 7 yes 19 yes 47 yes 21 yes 7 yes

Janson 12 yes 41 yes 47 yes 35 yes 20 yes

Melior 9 yes 12 yes 47 yes 16 yes 17 yes

Memphis 9 yes 3 yes 47 yes 6 yes 4 yes

Modern 6 yes 41 yes 47 yes 35 yes 20 yes

News No. 9 9 yes 41 yes 47 yes 35 yes 20 yes

Palatino 3 yes 12 yes 47 yes 21 yes 17 yes

Plantin 7 yes 19 yes 47 yes 35 yes 17 yes

Raleigh 9 yes 10 yes 47 yes 16 yes 17 yes

Rockwell * no 12 yes 47 yes 4 no 4 yes

Sabon 7 yes 19 yes 47 yes 35 yes 17 yes

Schneidler 7 yes 18 yes 11 yes 35 yes 17 yes

Serif Gothic Light 12 yes 2 no 47 yes 21 yes 2 no

Souvenir 1 yes 14 no 11 yes 35 yes 3 yes

4 yes Stymie Light 9 yes 12 yes 47 yes 6 yes

4 no Tiffany 8 yes 41 yes 11 yes 35 no

17 yes Times Bold 13 yes 41 yes 47 yes 35 yes

13 yes 17 yes Times Roman 12 yes 41 yes 47 yes

21 yes 8 yes Trump Medieval 6 yes 11 yes 47 yes 14 yes 20 yes Windsor Light 9 yes 41 yes 47 yes 209

TABLE 1 (Continued)

SEQUENCE

UNIDENTIFIED TYPEFACE figl fig 2 fig 3 fig 4 NAME

Search Match Search Match Search Match Search Match Search Match Yield Found? Yield Found? Yield Found? Yield Found? Yield Found?

Adminster Book 43 yes 4 yes 45 yes 3 yes 45 yes

American Typewriter 9 yes 5 yes 3 yes 34 yes 45 yes

Andover 43 yes 22 yes 45 yes 34 yes 45 yes

Artcraft Light 43 yes 22 yes 45 yes 1 yes 45 yes

Aurora 43 yes 25 yes 45 yes 34 yes 45 yes

Baskerville 43 yes 5 yes 45 yes 34 yes 45 yes

Baskerville Bold 43 yes 5 yes 45 yes 34 yes 45 yes

Baskerville II 43 yes 5 yes 45 yes 7 yes 45 yes

Bembo 43 yes 16 yes 45 yes 34 yes 45 yes

Bodoni 43 yes 25 yes 45 yes 3 yes 45 yes

Bodoni Book 9 yes 25 yes 45 yes 3 yes 45 yes

Bookman 43 yes 2 no 45 yes 34 yes 45 yes

Caledonia 43 yes 25 yes 45 yes 34 yes 45 yes

Cartier 43 yes 2 yes 10 yes 34 yes 4 yes

Caslon #3 43 yes 20 yes 45 yes 34 yes 45 yes

Caslon #224 43 yes 20 yes 45 yes 34 yes 45 yes

Century Oldstyle 43 yes 25 yes 45 yes 34 yes 45 yes

34 yes 45 yes Century Schoolbook 43 yes 25 yes 45 yes

34 yes 45 yes Century Textbook 43 yes 25 yes 45 yes

34 yes 45 yes Cheltenham Bold 43 yes 25 yes 45 yes

34 yes 45 yes Clarendon Book 43 yes 5 yes 45 yes

2 yes 45 yes Clearface Bold 43 yes 3 yes 45 yes

7 yes 45 yes Colonial 43 yes 18 yes 45 yes

34 yes 45 yes Congress Medium 43 yes 22 yes 45 yes

1 yes 45 yes Copperplate Gothic 43 yes 2 no 45 yes 210

TABLE 1 (Continued)

S E Q U E N C E UNIDENTIFIED TYPEFACE 8 1 fig 2 8 3 ;4 NAME

Search Match Search Match Search Match Search Match Search Match Yield Found? Yield Found? Yield Found? Yield Found? Yield Found?

Galliard Roman 9 yes 20 yes 45 yes 34 yes 45 yes

Garamond 43 yes 20 yes 10 yes 4 yes 45 yes

Garamond No. 3 43 yes 20 yes 10 yes 4 yes 45 yes

Garth Graphic 43 yes 20 yes 45 yes 7 yes 45 yes

Goudy 43 yes 20 yes 45 yes 7 yes 4 yes

Goudy Oldstyle 43 yes 20 yes 45 yes 7 yes 4 yes

Janson 43 yes 20 yes 10 yes 34 yes 45 yes

Melior 43 yes 20 yes 45 yes 34 yes 45 yes

Memphis 9 yes 22 yes 45 yes 34 yes 45 yes

Modern 43 yes 5 yes 45 yes 34 yes 45 yes

News No. 9 43 yes 25 yes 45 yes 34 yes 45 yes

Palatino 43 yes 4 yes 10 yes 34 yes 45 yes

Plantin 43 yes 22 yes 10 no 7 yes 45 yes

Raleigh 43 yes 4 yes 10 yes 34 yes 1 yes

Rockwell 9 yes 22 yes 10 no 34 yes 3 yes

Sabon 43 yes 22 yes 10 yes 34 yes 45 yes

yes 45 yes Schneidler 43 yes 4 yes 45 yes 34

34 no 45 yes Serif Gothic Light 43 yes 22 no 45 yes

34 yes 45 yes Souvenir 43 yes 25 yes 45 yes

34 yes 3 yes Stymie Light 9 yes 22 yes 45 yes

* * 45 yes no 45 yes Tiffany 43 yes no

yes 45 yes 34 yes 45 yes Times Bold 43 yes 20

yes 45 yes 34 yes 45 yes Times Roman 43 yes 22

yes 45 yes 34 yes 45 yes Trump Medieval 43 yes 22

yes 3 yes 34 yes 45 yes Windsor Light 43 yes 22 211

TABLE 2

Average Search Yield and Probability of Success for Each Sequence.

Number of Nseq Searches 'seq seq ^seq seq Sequence for which Number of yieldswere Total Average Number of Probability Searches recorded Yield Yield Successes of Success

A 50 48 980 20.4 48 96

C 50 50 977 19.5 50 100

G 50 49 1326 27.1 49 98

J 50 50 1786 35.7 50 100

L 50 50 1806 36.1 50 100

M 50 49 783 15.9 48 96

P 50 50 1592 32.1 50 100 Q 50 50 1540 30.8 50 100 R 50 49 993 20.3 49 98

T 50 50 884 17.7 50 100

W 50 49 389 7.9 49 98

a 50 50 1360 27.6 48 96

e 50 50 2137 41.8 50 100

f 50 50 1289 26.5 48 96

682 14.3 47 94 g 50 50 1 50 50 1946 38.9 50 100

2 50 49 823 17.3 46 92

3 50 50 1851 38.1 48 96

4 50 49 1253 24.9 48 96

50 50 1999 40.0 50 100 212

TABLE 3

Average Search Yield and Probability of Success For Any Search Based on One Sequence.

Ni Ti Yi Si Pi

Total Searches Sum of all Average Sum of all Probability of performed with search yields yield for successes for success with

one sequence for one searches all searches one sequence

sequence with one with one

sequence sequence

1000 26396 26.0 978 97.8 213

TABLE 4

Probability of Success of Searches Based on One or More Sequences

n Pn

No. of Sequences Probability on which of success search is based

1 97.80

2 95.65

3 93.54

4 91.45

5 89.47

6 87.54

7 85.58

8 83.70

9 81.86

10 81.85

11 78.29

12 76.57

13 74.89

14 73.24

15 71.63

16 70.05

17 68.51

18 67.00

19 65.53

20 64.08 214

TABLE 5

Average Search Yield of Searches Based on One or More Sequences.

n Yn

1 26

2 13

3 7

4 4

5 2

6 1

7 *

8 *

9 *

10 *

11 *

12 *

13 *

14 *

15 *

16 *

17 *

18 *

19 *

20 *

* = values less than 1 215

TABLE 6

Average Number of Sequences Needed to Narrow the List to One and the Probability of Success of TYPE-ID.

Average Ptype-id Number of Sequences of Needed to Probability success with Narrow the TYPE-ID List to One

6 85.58 216

Appendix G

TYPEFACE NOMENCLATURE 217

Appendix G

TYPEFACE NOMENCLATURE

The purpose of this appendix is to introduce and define some of the

typographic terms that may be encountered elsewhere in this paper.

Depending on which source is consulted, a single letter of type can be

"character," "sort," "element." "character" called a a or an While appears to

"sort" term for a be the most widely accepted term, the term (which is the

"element" single piece of metal type) is used occasionally. The term is most

often used to describe a portion of a character. A serif, for example, might

be considered an element.

called "capital There are two sets of characters. In the vernacular, these are

"uppercase" letters" letters." and and "small The printer calls these

"lowercase" calligrapher call them characters, respectively, while the may

"miniscules." "majuscules" most often employs the and The typographer

"caps" "lowercase," "uppercase" substitutes for terms and but sometimes

"miniscule" "majuscule" "uppercase." reserved The terms and are generally

Roman bookhands and are less for discussions oftypefaces based on early

discussion of Roman letterforms. Phillip Gaskell, however, frequently used in

Type" Nomenclature for the Letterforms of Roman in his article entitled "A

"miniscules." "capitals" and divides his characters, oddly, into

often appear to sit on an When characters are set in combination, they

"baseline." is referred to as the imaginary straight line. This line generally 218

The height to which caps extend above the baseline is called, generally, the "cap height."

The imaginary straight line drawn through the highest point of a cap and

line." parallel to the baseline is called the "cap

"x" The height to which the lowercase extends above the baseline is called

"x-height." the There are a number of terms that have been used to denote the imaginary line drawn through the highest point of the x and parallel to

line," line" line." the baseline. These terms include "mean "x and "waist

line" "Mean appears to be the most generally accepted term.

The usually vertical straight-line portions of such characters as the lowercase

"ascenders." b, d, f, h, k, and I which extend above the mean line are called

The imaginary line drawn through the highest point of an ascender and

line." line" parallel to the baseline is called the "ascender The ascender is a

line" distinct entity from the "cap but some sources neglect to make this distinction.

which The portions of such characters as the lowercase g, q, p, y, and j

"descenders." extend below the baseline are called The imaginary line drawn through the lowest point of a descender and parallel to the baseline

line." is called the "descender 219

There are often two different stroke thicknesses within a single character.

stroke," The thicker of stroke," these has been called the "thick "accented

stroke," "upstroke," "fat stroke," and and the thinner of these the "thin

stroke," "downstroke," "hairline." "lean and

A character which has only one stroke thickness is said to have "no

contrast," contrast," tone," "monotone." "geometric "evenness of or to be A

"contrast," character which has strokes of varying thickness is said to have

"modelling," "shading." or A small degree of stroke variation has been

contrast" contrast." termed "low or "gradual A great deal of stroke variation

contrast" contrast." has been termed "high or "abrupt

The arrangement of thick and thin strokes within a character (particularly

"o" round characters such as the or "c") can be used to determine the "axis

stress" of of a character.

stress" Characters whose axes of stress are vertical are said to have "vertical

shading." or "vertical A character whose axis of stress is not vertical is said to

"rounded," "oblique," "inclined," "angled," "diagonal" have or stress.

The vertical strokes of such characters as the cap E, I, M and others, have

strokes." "stems," stems" been called "main and "main A stem (or other

"oblique," structure) that is neither horizontal nor vertical may have an

"obtuse," "grave," "diagonal" "inclined," "angled," "acute," or orientation. 220

The strokes that are finishing commonly found at the ends of stems are

"serifs." called A number of different kinds of serifs have been identified.

those who Indeed, attempt to classify typefaces regard serif structures as an

criterian important for classification. There are a number of synonyms for

each type of serif.

"Bodoni" The thin, horizontal serifs that are found on such typefaces as have

"thin-line," "straight," "hairline," "Modern" been called or serifs.

"Stymie" The thick serifs which can be found on such typefaces as or

"Memphis" and which are often as thick as the stems to which they are

"slab," "square," "block," "flat," "Clarendon," attached, have been called

"Egyptian," "Antique" or serifs.

The structure which softens of fills in the angle between a serif and the stem

"bracket," "cove," "fillet," to which it is attached has been called a a a and a

"radius."

"triangular," "wedge," "hybrid" Triangular serifs have been called and serifs.

The serif-like formations on the ends of the arms of some characters, such

"beaks," serifs,""spurs," "half- as the cap T and E have been called "wing

serifs," "barbs." and 221

The ball-shaped terminal on such characters as the lowercase g, r, and f has

"ear," dot," "hook," "lug," terminal." been called an "tail and "ball When this

"tear-drop," "bulb." same terminal is pear-shaped, it has been called a or

The enclosed areas of such characters as the cap O, Q, and the lowercase o and the partially enclosed areas of such characters as the cap H, are called

"counters."

The rounded stroke of such characters as the lowercase b, c, and p is called,

"bowl." generally, the

The junction at the upper-most portions of such characters as the cap A, W,

"apex." and M is called an The junction at the lower-most portions of such

"vertex." characters as the cap V, and W is called a

The piece which connects the upper bowl of the lowercase g to the lower

"link," "neck," element." bowl has been called the and "joining

near the baseline be The bottom portions of stems that may be found may

"feet." called

characters as the E and T are called The horizontal strokes of such cap 222

The cross-piece of such characters as the cap A and lowercase e has been

"bar." called a

The protrusions found at the lower-right regions of the cap G are called

"spurs" "nibs." or

"hook." The curved portion of the J may be called a 223

Appendix G

NOTES

Type," 1. Philip Gaskell, "A Nomenclature for the Letterforms of Roman Visible Language (Winter 1976). 41-61. 224

Appendix H

TYPEFACE CLASSIFICATION SYSTEMS 225

Appendix H

TYPEFACE CLASSIFICATION SYSTEMS

Since the beginning of the twentieth century there have been devised a number of different typeface classification systems. Below is presented, in chronological order by date of publication, a representative sampling.

Theodore Low De Vinne (1900)1

functional" Devinne constructed a "structural and but not historical classification of typefaces. There are nine categories:

1. Roman 2. Italic 3. Script 4. Blackletter 5. Gothic 6. Italian 7. Title or Fat Face 8. Antique 9. Ornamentals

F. Thibaudeau (1903)2

for Thibaudeau determined that serif structures are the best basis classification. His four main divisions are:

1. Antique (Sans Serif) 2. Egyptienne (Egyptian Serif) 226

3. Romain Elzevir (Old Style Serif) 4. Romain Didot (Modern Serif)

Legros & Grant (1922)3

According to Legros and Grant, typefaces may be divided into three main groups. These are:

1. Old-style Faces 2. Modern Faces 3. Fancy Faces.

These main groups are further divided into 12 classes (Legros and Grant do

each class is derived). The not, however, specify from which main group categories are:

1. Black 2. Black Ecclesiastical 3. Sans Serif 4. Old Face 5. Old-Style Antique 6. DeVinne 7. Blackfriars 8. Cheltenham 9. Latin 10. Modernized Old-style 11. Modern 12. Antique 227

Beatrice Warde (1935)4

Beatrice Warde proposed a system based on the morphology of typefaces.

The physical shape of the characters, and not their lineages, is the sole

"old-face" criterian. For example, the widely used term was in Warde's

Group." ("Aldine" system called "a bias-stress roman type of the Aldine refers to type based on or produced by the Aldine Press, which under Aldus

Manutius [whose motto was "make haste slowly"] produced some of the most beloved printing of the incunabula period of printing). The specifics of this system need not be illustrated here.

lohnson & Berry (1953)

The classic specimen book Encyclopedia of Type FacesS arranges typefaces by an 18-category classification system:

1. Venetian Romans 2. The Old Faces 3. Eighteen-Century Transitional Types 4. Modern Faces 5. Twentieth-Century Faces 6. Calligraphic Types 7. Display Types 8. Private Press Types 9. Sans Serif Faces 10. Modified Sans Serif 11. Fat Faces and Heavy Display 12. Eqyptian or Antique Faces 13. Modified Egyptian Faces 14. Black Letter Types 15. Outline and Shaded Letters 16. Decorated Types 17. Three-Dimensional Faces 18. Script Types 228

Maximilien Vox (1954)6

System" The "Vox divides typefaces into nine groups (although in the original system there were 10). This system was highly acclaimed and other systems have been based, at least to some degree, on it. Vox said that the division names "were chosen for easy pronunciation and translation but they could easily be replaced by numbers. Here are the nine groups:

1. Manuaires (drawn letters) 2. Humanes (15th-century Romans) 3. Garaldes (15th-century French types) 4. Reales (Transitionals) 5. Didones (Moderns) 6. Incises (chiseled or wedge-shaped serifs) 7. Lineales (sans serifs) 8. Mecanes (Egyptians, slab serifs) 9. Scriptes (calligraphic types)

Deutsche Industrie Norman (DIN) System (1959)?

proposal The DIN System is considered by some to be the most complete

the categories and sub for the classification of typefaces. Here are categories:

1. ROMAN LETTER FORMS

1.1 Renaissance

bar 1.11 Early Form: e with oblique bar 1 12 Late Form: e with horizontal since 1890 1 13 Modern Form: types produced 229

1.2 Baroque

1.21 Dutch Form 1.22 English Form 1.23 French Form 1.24 Modern Form

1.3 Neo-CIassic

1.31 Early Form 1.32 Late Form 1.33 Newspaper Form 1.34 Modern Form

1.4 Free Roman

1.41 Victorian Form 1.42 Serifless Roman 1.43 Individual Form

1.5 Linear Romans

1.51 Early Form 1.52 Modern Form

1.6 Block Roman

1.61 Early Form 1.62 Late Form (Clarendon) 1.63 Modern Form 1.64 Typewriter Types

1.7 Roman Script

1.71 Broad pen 1.72 Flexible, Pointed Pen 1.73 Strokes of equal thickness 1.74 Brush Script

2. FRAKTUR LETTER FORMS

2.1 Textura

2.2 Rotunda

2.3 Schwabacher 230

2.4 Fraktur

2.5 Cursive

3. EXOTICS

3.1 Greek

3.2 Cyrillic

3.3 Hebrew

3.4 Arabic

3.5 Other

ATypI (1961)8

Typefounders, composing machine manufacturers, typographers, and printers from the United States and Europe were represented in the 1961

Association Typographique Internationale (ATypI) which recommended the following typeface classification system (the 10 classes presented here closely resemble those suggested by Vox):

1. Humane (15th-century Romans) 2. Geralde (15th-century French Types) 3. Reale (Transitionals) 4. Didone (Modern) 5. Mecane (Square Serifs) 6. Lineale (Sans Serif) 7. Incise (Wedge-serifs) 8. Manuaire (Display Types) 9. Scripte (Cursives) 10. Fractura (Blackletters) 231

The British Standards System (1965)9

This system was suggested by the British Standards Institution. There are

who believe that his system many should be universally adopted. It is similar to the Vox system.

1. Graphic (drawn letters) 2. Humanist (15th-century Roman types) 3. Geralde (15th-century French Types) 4. Transitional 5. Didone (Modern) 6. Lineale (Sans Serif) 7. Slab Serif 8. Glyphic (calligraphic) 9. Script

Alexander Lawson (1971)10

"rational" Alexander Lawson saw a need for a more system and so devised this eight-category classification:

1. Blackletter

2. Oldstyle

a. Venetian b.Aldine-French c. Dutch-English

3. Transitional

4. Modern

5. Square Serif 232

6. Sans Serif

7. Script-Cursive

8. Display-Decorative

Marshall Lee (1979)

In Bookmaking,vi Marshall Lee divides typefaces into the following general

categories and sub categories:

1. Roman

a. Old Style b. Transiltional b. Modern d. Egyptian e. Miscellaneous

2. Abstract

a. Sans Serif b. Block Serif

3. Cursive

a. Italics b. Scripts

4. Decorative (& Blackletters)

Rookledge & Perfect (1983)

Gordon Rookledge and Christopher Perfect in Rookledge's International

new" Typefinder.^ present an "entirely classification system which is based

features." on 'the grouping of typefaces according to specific design While the names of the classes are new, the system is loosely based on the British 233

Standards System. Of particular interest is the treatment given to decorative and display faces. Rather than lumping them all together in one

"Fancy" miscellaneous category (eg. Legros and Grant's types, DeVinne's

"Ornamentals," etc.) they are subdivided into a number of classes and subclasses (including one called "Modified Outrageous"). Of further interest is the fact that some typefaces appear in more than one category. To quote from the source: "There is a small overlap between the Text and Decorative parts of the book. For instance, some Decorative typefaces in special

vice-versa." circumstances may be used for continuous text setting and Here are the categories and sub categories:

A. TEXT TYPES

1. Sloping e-bar 2. Angled Stress/Oblique Serifs 3. Vertical Stress/Oblique Serifs 4. Vertical Stress/ Straight Serifs 5. Abrupt Contrast/Straight Serifs 6. Slab Serif 7. Wedge Serif 8. Sans Serif

B. DISPLAY FACES

1. Flowing Scripts 2. Non-Flowing Scripts 3. Unmodified 4. Fat and Thin Face 5. Ornamental 6. Modified Serif 7. Modified Sans Serif 8. Modified Outrageous 234

Martin Solomon (1986)

In the Art of Typography^ Martin Solomon identifies the following "basic

styles" of type:

1. Roman

a. Old Style b. Transitional c Modern d. Primer (or book) e. Fat Face f. Egyptian (Antique) g. Italic

2. Script

a. Gothic Scripts b. Latin Scripts

3. Gothic

4. Ornamental

5. Period

Archibald Provan (1989)

Rochester Institute of Technology Professor Archibald Provan has proposed the following "Typeface Design Grouping Scheme"14 to the International

structure" Standards Organization. It consists of a "three-level hierarchical in which similar or substitutable typefaces are grouped together. This system is perhaps the most comprehensive yet devised. 235

1. Uncials

1.1 Single Alphabet

1.1.1 sans serif 1.1.2 serif

1.2 Duplex Alphabet

1.2.1 sans serif 1.2.2 serif

2. Inscriptionals

2.1 Solids

2.1.1 sans serif 2.1.2 serif

2.2 Inlines

2.2.1 sans serif 2.2.2 serif

2.3 Outlines

2.3.1 sans serif 2.3.2 serif

3. Blackletters

3.1 Formal Style

3.1.1 sans serif 3.1.2 serif 3.1.3 sans serif engraved 3.1.4 serif, engraved

3.2 Round Style

3.2.2 serif 3.2.3 sans serif, engraved 3.2.4 serif, engraved 236

3.3 Hybrid

3.3.1 sans serif 3.3.2 serif 3.3.3 sans serif 3.3.4 serif, engraved

3.4 Informal Style

3.4.1 sans serif 3.4.2 serif 3.4.3 sans serif, engraved 3.4.4 serif, engraved

4. Serifs

4.1 Oldstyle

4.1.1 Venetian 4.1.2 Garalde 4.1.3 Dutch/English

4.2 Transitional

4.2.1 Direct line 4.2.2 Modified

4.3 Modern

4.3.1 Italian 4.3.2 Fat Face

4.4 Contemporary

4.4.1 Eclectic 4.4.2 Fine serif 4.4.3 Lettering

4.5 Legibility

4.5.1 Rounded (traditional) 4.5.2 Square (modern)

4.6 Square Serif

4.6.1 Monotone 4.6.2 Clarendon 4.6.3 French Clarendon 4.6.4 Short (stub) 237

4.6.5 Typewriter 4.6.6 Dot Matrix

4.7 Latin

4.7.1 Solid 4.7.2 inline

4.8 Engraving

4.8.1 Barbed Serif 4.8.2 Straight Serif (fine)

4.9 Art Nouveau

4.9.1 French

4.10 Computer

4.10.1 OCR 4.10.2 Digital

4.11 Miscellaneous

5. Sans Serif

5.1 Gothic

5.1.1 Grotesque 5.1.2 Neo-grotesque 5.1.3 Typewriter

5.2 Humanist

5.2.1 Classical 5.2.2 Non-classical 5.2.3 Typewriter

5.3 Stress Variation

5.3.1 Broad Pen 5.3.2 Casual 5.3.3 Typewriter

5.4 Art Deco

5.4.1 Standard 5.4.2 Modified 5.4.3 Thin Line 238

5.5 Geometric

5.5.1 Round 5.5.2 Elliptical 5.5.3 Stylized 5.5.4 Typewriter

5.6 Computer

5.6.1 OCR 5.6.2 Digital

5.7 Art Nouveau

5.7.1 French

6. Scripts

6.1 Joined

6.1.1 Formal 6.1.2 Informal 6.1.3 Monotone

6.2 Unjoined

6.2.1 Formal 6.2.2 Informal 6.2.3 Monotone 6.2.4 Brush 6.2.5 Cursive 6.2.6 Calligraphic 6.2.7 Ronde

7. Ornamentals

7.1 Inline

7.1.1 Sans Serif 7.1.2 Serif

7.2 Outline

7.2.1 Sans Serif 7.2.2 Serif 239

7.3 Decorative

7.3.1 Sans Serif 7.3.2 Serif

7.4 Three-Dimensional

7.4.1 Sans Serif 7.4.2 Serif

7.5 Tuscan

7.5.2 Serif

7.6 Stencil

7.6. 1 Sans Serif 7.6.2 Serif

7.7 Reversed

7.7.1 Sans Serif 7.7.2 Serif

7.8 Engraved

7.8.1 Sans Serif 7.8.2 Serif

7.9 Miscellaneous

7.9.1 Sans Serif 7.9.2 Serif 240

Appendix H

NOTES

1. Typefaces," James Mosley, "New Approaches to the Classification of The British Printer (March 1960). 92.

2. Ibid. 92.

3. L. A. Legros and J. C. Grant, Typographical Printing Surfaces: The Technology and Mechanism of their Production (London, Longman's Green & Co., 1916). 82-4.

4. Mosley, Op. Cit. 93.

5. A. F. Johnson, The Encyclopedia of Typefaces (London: Blandford Press, 1953).

Typefaces." 6. Maximilien Vox Devises a New Classification of Printing World (July, 30, 1954). 120-21.

7. Mosley, Op. Cit. 94.

8. Alexander Lawson, Printing Types: An Introduction (Boston, Mass.: Beacon Press. 1971). 37.

9. Ibid. 38.

10. Ibid. 47.

11. Marshall Lee, Bookmaking, 2nd Ed. (New York: R. R. Bowker Co., 1979). 74-6.

12. Christopher Perfect and Gordon Rookledge, Rookledge's International Typefinder (New York: Frederick C. Beil, 1983). 5.

13 Martin Solomon, The Art Of Typography (New York: Watson-Guptill Publications, 1986). 64-77.

Scheme," 14. Archibald Provan, "Proposed Typeface Grouping ISO/DIS 9541. 241

Appendix I

TYPEFACE IDENTIFICATION SYSTEMS 242

Appendix I

TYPEFACE IDENTIFICATION SYSTEMS

A number of articles and books have been written on the subject of

typeface identification. Some present so-called typeface identification

systems. A representative sampling of these systems ispresented below.

All of these systems have one thing in common: they are essentially useless.

The system which "potential" only was found to have any was the "Panose

System" which is the subject of Chapter 2.

The gal Method

"gal" The method, as presented by K. A. Kok in Print in Britain^, is an

identification system based on the three lowercase letters g, a, and I. More specifically, from the upper bowl and ear of the lowercase g, the upper curve and terminal of the lowercase a and the top serif of the lowercase I, the user is supposed to be able to identify a typeface.

system," "In order to user the Kok writes, "it is necessary to have a sharp

this article." The print, a magnifying glass, and the illustrations accompanying

glass" print" "sharp is the unidentified typeface specimen, the "magnifying is

"illustrations" and the are to for close scrutiny of the unidentified specimen, be used for comparison. 243

To a typeface the gal identify using method the user must: (1) determine

which of the supplied gal-method illustrations most closely resembles his

unidentified sample and record the numbers assigned to those illustrations,

and look (2) up those numbers in the supplied index.

For the user to determine that the commercial name of this unidentified

"Bodoni" typeface is he would have to: (1) conclude, by comparison

between his typeface and the illustrations, that his typeface has a lowercase

"g" similar to illustration "a" 38, a lowercase similar to illustration 1, and a

"I" lowercase similar to illustration 4; and (2) look in the index for a typeface

with g-38, a-1, and d-4.

To test the gal method, two subjects were asked to identify 5 typefaces each. Both subjects concluded that the system is useless and gave up in frustration.

The principal problem with the gal method is that many of the supplied illustrations are so similar that it is difficult to distinguish between them.

Problems of comparison are compounded when the user's unidentified sample is of a different point size than the illustrations. Since it is essential

"correct" that the user choose the illustration, and since this is an essentially impossible task, the system is useless. 244

Rookledge's Earmark Tables

In Typefinder,2 Rookledge's International Gordon Rookledge and Christopher

"earmark" Perfect present their typographical (ie., "distinctive identifying feature") tables. This system is similar to the afore-mentioned gal method in that the user must make visual comparisons between characters from his unidentified sample and supplied illustrations. Here, however, any letter of the alphabet, both uppercase and lowercase, plus figures and ampersands, can be used for comparison.

The three elements of the earmark tables are: (1) the unidentified typeface;

(2) the supplied earmark table illustrations (there is a separate table for each letter, and figure and one for ampersands); and (3) the supplied typeface

"major" specimen index (where full font layouts of every text typeface are supposedly displayed).

To identify a typeface using the earmark tables the user must: (1) decide which character he wishes to use for comparison, (2) determine which of the supplied illustrations in the earmark table of that chosen character most

record the specimen numbers closely resembles his unidentified sample, (3)

specimens in the assigned to that supplied illustration, and (4) look up the

sample. supplied index and compare them to his

commercial name of this unidentified For the user to determine that the

"Meridian" chosen the A as his letter typeface is (and assuming he has cap 245

for comparison) he must: scan (1) the 47 supplied illustrations of the cap

until "earmarks" he finds one which appears to have the same as his

unidentified sample, (2) determine that supplied illustration #35, which

represents specimens 218, 219, 233, and 257, is the one which most closely

resembles his sample, (4) look up specimens 218, 219, 233, and 257 in the

supplied specimen index and visually compare them to his sample, and (5)

determine that specimen 233, Meridian, is the specimen which matches his

unidentified sample.

tables' To test the earmark effectiveness, 2 subjects were asked to try to

identify 5 typefaces each. Both subjects concluded that the system is

impossible to use and gave up in frustration.

The principal problems with the earmark tables are:

(1) a single earmark table illustration may represent a number of different

typefaces. For example, cap M illustration #4 supposedly represents the cap

M's of 67 different typefaces. The user who determines that the cap M of his

sample is best represented by this illustration #4 has an awesome task

- of all 67 ahead he may have to look in the supplied index at specimens typefaces before he can be certain he has found a match.

the earmark tables do (2) The user is required to interpolate. Again, because

representative not present illustrations of every typeface, but rather samples, 246

the user is guess" required, in most instances, to make his "best as to which illustration most nearly resembles his sample.

(3) Every major typeface is not represented by the earmark table illustrations. For example, none of the 47 illustrations of the cap A and none

"Palatino." of the 44 illustrations of the cap B represents the typeface

(4) The earmark tables contain errors. For example: (a) Cap M illustration #4 supposedly represents 67 typefaces all of which are supposed to have,

"pointed" among other things, vertexes. However, a search of the specimen

"flat" index revealed that a number of the 67 have or other kinds of vertexes; (b) Cap E illustration #27 represents 30 typefaces all of which are

"long" supposed to have bottom arms. After looking at only a few of the 30 it became clear that not all of the examples could be clearly identified as having long arms. Furthermore, there are a number of other typefaces, not included in the 30, that do have long arms.

Because of the above problems, the earmark tables cannot be used

identification system. effectively as a typeface

The 100-ems Method3

follow a procedure to the user of the 100-ems method must two-step

count" determine the "average character identify a typeface. First, he must 247

per 100 ems of [ACC] his typeface, and second, he must consult a chart in which typefaces are listed according to this count.

The user may determine the ACC of his unidentified typeface by solving the

= equation ACC A/M x 100 where A is the total number of characters in the

unidentified sample and M is the total measure in ems of the sample.

No testing was necessary to conclude that this system has the following

- time- problems: (1) The user is required to count characters a possiblly

consuming task; (2) the sample size of the unidentified typeface must be

large enough for the user to obtain an accurate ACC - small samples of

type, such as headlines, sub-heads, or short paragraphs, are difficult to

gauge and therefore cannot be identified with certainty; (3) Today's

typesetters, using digital equipment, routinely alter both the set-width and

letterfit of characters. Therefore, the ACC per 100 ems of a given typeface

may vary.

A further problem with this system is that the user is not required to

to examine individual characters of type and is therefore not taught identify,

of specific typefaces on his own, the characteristics and peculiarities

attempts to the (similarly afflicted would be the musician who identify

the average number of composer of a musical composition by calculating

notes per measure). 248

The 100-ems method (or any system based on an average character count, such as the character-per-pica count) is not a reliable typeface identification procedure.

The Panose System4

The Panose system is discussed in Chapter 2. 249

Appendix I

NOTES

"gal," 1.K. A. Kok, Print in Britain (date unknown). 430-433.

2. Christopher Perfect and Gordon Rookledge, Rookledge's International Typefinder (New York: Frederic C. Beil, 1983). 97-153.

3. Kok, Op. Cit. 430-433.

4. Benjamin Bauermeister, A Manual of Comparative Typography: The Panose System (New York: Van Nostrand Reinhold Col, 1988). 250

Appendix J

TYPE-ID DESIGNS 251

Appendix J

TYPE-ID DESIGNS

The following are design proposals for various elements of TYPE-ID.

Included herein are a cover design and some profile-generator-question page designs. All were produced on the Xerox 6085 Documenter.

Collectively, they, and some of the page designs included in the thesis proper, were awarded a 1989 XEPA -- Xerox Electronic Publishing Award for design. 252

CATEGORY # 1

SERIF TYPEFACES 253

a. pointed APEX -1 What kind of Apex does the A b. flat Cap A have? A THE APEX is the portion of the Cap A where the two c. angled main stems meet (see illustration above). The most common apex among the serif typefaces is the pointed apex. It is found on such typefaces as Bodoni and If Baskerville. you were to take a pair of scissors and the off snip top a pointed apex, you would be with A left "sheared" a apex. There are two types of sheared d. rounded apexes: sheared flat (ie., horizontal or parallel to the baseline) and sheared angled (anything other than flat). A flat apex can be found on such typefaces as ITC Bookman and Melior. An angled apex can be found on Garamond (Stempel and and A Berthold) ITC Zapf

International. Rounded apexes can be found on e. concave Parsons and Romic. ITC Souvenir is considered for

purposes here to have a pointed apex. There are some typefaces, such as Tiffany and Caslon, which have apexes that appear to have been hollowed out. These we call concave apexes. A Some typefaces have serif structures at their apexes. f. full serif Rockwell, for instances, has a full serif at its apex, while Lubalin and Beton have half-serifs.

4 A

g. half-serif left A

h. half-serif right AN

v *. VUw-o -. o O^'SsV! -Ss^xv! 254

APEX Is the Apex Stepped?

THE APEX is the top portion of the Cap A where the two main stems meet (see illustration above). The apex

is stepped if one stem extends beyond the other (examples: Caslon, Bernhard Modern). The apex is joined if both stems meet and end at the same point (examples: Baskerville, Times Roman). Seriffed

"unusual" apexes or apexes belong to the other category (examples: Rockwell, Memphis).

4& 255

A-3

a. horizontal What is the orientation A of the bar b. curved of the A? Cap BAR AA

THE BAR of the Cap A is the cross-piece that runs c. angled between the two main stems (indicated in above diagram). The Cap A's of most seriffed typefaces have horizontal ( ie., straight and parallel to the baseline) bars (examples: Baskerville, Bodoni, Times Roman, et A al.) A bar that has a smooth, single curve is called a curved bar (examples ITC Benguiat, Galliard). Bars d. other which are straight but not horizontal, whatever their orientations, are called angled bars (example: Fontana). Bars which have more than one curve or which are unusual (or absent) belong to the other A category (example: Lo-type).

?*.>

mm